Articles Safety and immunogenicity of the protein-based PHH-1V compared to BNT162b2 as a heterologous SARS-CoV-2 booster vaccine in adults vaccinated against COVID-19: a multicentre, randomised, double-blind, non-inferiority phase IIb trial Júlia Corominas,a,aa Carme Garriga,a,aa Antoni Prenafeta,a Alexandra Moros,a Manuel Cañete,a Antonio Barreiro,a Luis González-González,a Laia Madrenas,a Irina Güell,a Bonaventura Clotet,b,r Nuria Izquierdo-Useros,b,o Dàlia Raïch-Regué,b Marçal Gallemí,b Julià Blanco,b,o,r,s Edwards Pradenas,b Benjamin Trinité,b,o Julia G. Prado,b,o,s Oscar Blanch-Lombarte,b Raúl Pérez-Caballero,b Montserrat Plana,c,o Ignasi Esteban,c Carmen Pastor-Quiñones,c Xavier Núñez-Costa,d,y,z Rachel Abu Taleb,d,y,z Paula McSkimming,d,y,z Alex Soriano,e,o Jocelyn Nava,e Jesse Omar Anagua,e Rafel Ramos,f ,p Ruth Martí Lluch,f ,p Aida Corpes Comes,f Susana Otero Romero,g,q,v Xavier Martinez Gomez,g,q Carla Sans-Pola,g,w,x José Moltó,h,o Susana Benet,h Lucía Bailón,h Jose R. Arribas,i,o Alberto M. Borobia,i,o,t Javier Queiruga Parada,i,t Jorge Navarro-Pérez,j Maria José Forner Giner,j Rafael Ortí Lucas,j María del Mar Vázquez Jiménez,k Salvador Oña Compán,k Melchor Alvarez-Mon,l Daniel Troncoso,l Eunate Arana-Arri,m Susana Meijide,m Natale Imaz-Ayo,m Patricia Muñoz García,n,u Sofía de la Villa Martínez,n Sara Rodríguez Fernández,n Teresa Prat,a,∗ Èlia Torroella,a and Laura Ferrera a HIPRA, Avinguda de la Selva, 135, 17170, Amer, Girona, Spain b IrsiCaixa AIDS Research Institute, Carretera de Canyet, s/n, Can Ruti Campus, 08916, Badalona, Spain c AIDS Research Group, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), C/ del Rosselló, 149, 08036, Barcelona, Spain d Veristat, LLC, Barcelona, Spain e Hospital Clínic of Barcelona, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), C. de Villarroel, 170, 08036, Barcelona, Spain f Institut Universitari d’Investigació en Atenció Primària Jordi Gol (IDIAP Jordi Gol), Biomedical Research Institute, Girona (IdIBGi), Catalan Institute of Health, Carrer del Dr. Castany, s/n, 17190, Salt, Girona, Spain g Hospital Universitari Vall d’Hebron, Pg. de la Vall d’Hebron, 119, 08035, Barcelona, Spain h Infectious Diseases Department, Hospital Universitari Germans Trias i Pujol, Carretera de Canyet, s/n, Badalona, Spain i Infectious Diseases Unit, Internal Medicine Department, La Paz University Hospital, IdiPAZ, C. de Pedro Rico, 6, 28029, Madrid, Spain j Hospital Clínico Universitario Valencia, Av. de Blasco Ibáñez, 17, 46010, València, Spain k Hospital Regional Universitario de Málaga, Av. de Carlos Haya, 84, 29010, Málaga, Spain l Hospital Universitario Príncipe de Asturias, Av. Principal de la Universidad, s/n, 28805, Alcalá de Henares, Madrid, Spain m Scientific Coordination, Biocruces Bizkaia HRI, Osakidetza, Cruces Plaza, 48903, Barakaldo, Bizkaia, Spain n Instituto de Investigación Sanitaria Hospital Gregorio Marañón, C. del Dr. Esquerdo, 46, 28007, Madrid, Spain o CIBER Enfermedades Infecciosas (CIBERINFEC), Instituto de Salud Carlos III, Av. de Monforte de Lemos, 5, 28029, Madrid, Spain p Department of Medical Sciences, School of Medicine, University of Girona, Plaça de Sant Domènec, 3, 17004, Girona, Spain q Unitat Docent Vall d’Hebron, Universitat Autònoma de Barcelona, Plaça Cívica, 08193, Bellaterra, Barcelona, Spain r Infectious Diseases and Immunity, Faculty of Medicine, Universitat de Vic-Universitat Central de Catalunya (UVic-UCC), Carrer de la Sagrada Família, 7, 08500, Vic, Spain s Germans Trias i Pujol Research Institute (IGTP), Carretera de Canyet, s/n, Badalona, Spain t Spanish Clinical Research Network - SCReN, Spain u CIBER Enfermedades Respiratorias- CIBERES (CB06/06/0058), Madrid, Spain v Department of Neurology/Neuroimmunology, Centro de Esclerosis Múltiple de Catalunya (Cemcat), Hospital Universitari Vall d’Hebron, Vall d’Hebron Barcelona Hospital Campus, Pg. de la Vall d’Hebron, 119, 08035, Barcelona, Spain w Department of Pharmacology, Therapeutics and Toxicology, Universitat Autònoma de Barcelona, Plaça Cívica, 08193, Bellaterra, Barcelona, Spain x Clinical Pharmacology Research Group, Vall d’Hebron Institut de Recerca, Pg. de la Vall d’Hebron, 119, 08035, Barcelona, Spain y Veristat, LLC, Toronto, Canada z Veristat, LLC, Pickmere, UK DOI of original article: https://doi.org/10.1016/j.lanepe.2023.100623 *Corresponding author. E-mail address: teresa.prat@hipra.com (T. Prat). aa Contributed equally and are joint first authors. www.thelancet.com Vol 28 May, 2023 1 Articles The Lancet Regional Summary Health - Europe Background A SARS-CoV-2 protein-based heterodimer vaccine, PHH-1V, has been shown to be safe and well- 2023;28: 100613 tolerated in healthy young adults in a first-in-human, Phase I/IIa study dose-escalation trial. Here, we report the Published Online 14 April interim results of the Phase IIb HH-2, where the immunogenicity and safety of a heterologous booster with 2023 PHH-1V is assessed versus a homologous booster with BNT162b2 at 14, 28 and 98 days after vaccine administration. https://doi.org/10. 1016/j.lanepe.2023. 100613 Methods The HH-2 study is an ongoing multicentre, randomised, active-controlled, double-blind, non-inferiority Phase IIb trial, where participants 18 years or older who had received two doses of BNT162b2 were randomly assigned in a 2:1 ratio to receive a booster dose of vaccine—either heterologous (PHH-1V group) or homologous (BNT162b2 group)—in 10 centres in Spain. Eligible subjects were allocated to treatment stratified by age group (18–64 versus ≥65 years) with approximately 10% of the sample enrolled in the older age group. The primary endpoints were humoral immunogenicity measured by changes in levels of neutralizing antibodies (PBNA) against the ancestral Wuhan-Hu-1 strain after the PHH-1V or the BNT162b2 boost, and the safety and tolerability of PHH-1V as a boost. The secondary endpoints were to compare changes in levels of neutralizing antibodies against different variants of SARS-CoV-2 and the T-cell responses towards the SARS-CoV-2 spike glycoprotein peptides. The exploratory endpoint was to assess the number of subjects with SARS-CoV-2 infections ≥14 days after PHH-1V booster. This study is ongoing and is registered with ClinicalTrials.gov, NCT05142553. Findings From 15 November 2021, 782 adults were randomly assigned to PHH-1V (n = 522) or BNT162b2 (n = 260) boost vaccine groups. The geometric mean titre (GMT) ratio of neutralizing antibodies on days 14, 28 and 98, shown as BNT162b2 active control versus PHH-1V, was, respectively, 1.68 (p < 0.0001), 1.31 (p = 0.0007) and 0.86 (p = 0.40) for the ancestral Wuhan-Hu-1 strain; 0.62 (p < 0.0001), 0.65 (p < 0.0001) and 0.56 (p = 0.003) for the Beta variant; 1.01 (p = 0.92), 0.88 (p = 0.11) and 0.52 (p = 0.0003) for the Delta variant; and 0.59 (p ≤ 0.0001), 0.66 (p < 0.0001) and 0.57 (p = 0.0028) for the Omicron BA.1 variant. Additionally, PHH-1V as a booster dose induced a significant increase of CD4+ and CD8+ T-cells expressing IFN-γ on day 14. There were 458 participants who experienced at least one adverse event (89.3%) in the PHH-1V and 238 (94.4%) in the BNT162b2 group. The most frequent adverse events were injection site pain (79.7% and 89.3%), fatigue (27.5% and 42.1%) and headache (31.2 and 40.1%) for the PHH-1V and the BNT162b2 groups, respectively. A total of 52 COVID-19 cases occurred from day 14 post-vaccination (10.14%) for the PHH-1V group and 30 (11.90%) for the BNT162b2 group (p = 0.45), and none of the subjects developed severe COVID-19. Interpretation Our interim results from the Phase IIb HH-2 trial show that PHH-1V as a heterologous booster vaccine, when compared to BNT162b2, although it does not reach a non-inferior neutralizing antibody response against the Wuhan-Hu-1 strain at days 14 and 28 after vaccination, it does so at day 98. PHH-1V as a heterologous booster elicits a superior neutralizing antibody response against the previous circulating Beta and the currently circulating Omicron BA.1 SARS-CoV-2 variants in all time points assessed, and for the Delta variant on day 98 as well. Moreover, the PHH-1V boost also induces a strong and balanced T-cell response. Concerning the safety profile, subjects in the PHH-1V group report significantly fewer adverse events than those in the BNT162b2 group, most of mild intensity, and both vaccine groups present comparable COVID-19 breakthrough cases, none of them severe. Funding HIPRA SCIENTIFIC, S.L.U. Copyright © 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Keywords: SARS-CoV-2; Protein-based vaccine; Heterologous boost; SARS-CoV-2 variants; Neutralizing antibodies Introduction the host cells and generates a neutralizing antibody The severe acute respiratory syndrome coronavirus 2 response.1,2 There is a positive correlation between the (SARS-CoV-2) identified in Wuhan, China, in December levels of SARS-CoV-2 neutralizing antibodies induced by 2019, has brought an urgency to prophylactic measures vaccination and the conferred level of protection, and globally. Consequently, the world has witnessed a strik- antibody levels have been previously used as endpoints ing change in the speed of vaccine development targeting for many viral vaccines.3 the SARS-CoV-2’s membrane spike protein, particularly Clinical trials for several approved SARS-CoV-2 vac- its receptor binding domain (RBD), which interacts with cines also revealed the induction of a rapid and specific the angiotensin converting enzyme 2 (ACE2) receptor in T-cell-mediated response in most participants noted 2 www.thelancet.com Vol 28 May, 2023 Articles Research in context Evidence before this study (available at covid19.trackvaccines.org) to check approved Despite the efforts of mass vaccination programs against and in development vaccines sorted by type. Consequently, SARS-CoV-2, the situation is far from being under control. we believe that PHH-1V, a SARS-CoV-2 vaccine based on a Only 68.5% of the world population, and, more starkly, fusion heterodimer consisting of the spike RBD sequence 24.6% in low-income countries have received at least one from the SARS-CoV-2 Beta (B.1.351) and Apha (B.1.1.7) dose of a COVID-19 vaccine. Moreover, because immunity variants, comes as a safe and efficient option for wanes over time after both vaccination and/or infection, heterologous boost administrations regimens considering booster immunisation is crucial to control the SARS-CoV-2 the interim data reported in this article for the HH-2 trial. infection rate and avoid saturation of health services. Added value of this study Accordingly, access to second-generation vaccines with a PHH-1V as a heterologous boost elicits a non-inferior broader variants scope and longer-lasting protection is neutralizing antibody response to SARS-CoV-2 previous and required. To date, NVX-CoV2373 (Novavax) is the only currently circulating variants when compared to the recombinant spike protein-based vaccine approved by WHO BNT162b2 boost. Noteworthy, PHH-1V conferred a superior and EMA, and more vaccines platforms are needed. One neutralizing antibody response against the Omicron variant of would consider recombinant protein-based to be concern (sublineage BA.1) at 14-, 28- and 98-days post- advantageous SARS-CoV-2 vaccines presentations; they are vaccination. PHH-1V as a booster dose induced a significant stable at 2–8 ◦ C, with good safety and stability profiles, and increase of CD4+ and CD8+ T-cells expressing IFN-γ, suitable for people with compromised immune systems. We considered a biomarker of successful vaccination resulting in also searched PubMed on July 11th, 2022, with no date or the development of effective humoral response. The language restrictions, and using the search terms “SARS- vaccination with PHH-1V did not induce a Th2-biased CoV-2” and “protein-vaccine” in the Title/Abstract. Our immune response, an important aspect previously associated search revealed the vaccines RelCoVax®, S-26019-b, CoV2 with ineffective vaccines that induce vaccine-associated preS dTM-AS03, zf2001, CoVaccine HT, NARUVAX-C19, enhanced respiratory disease. The PHH-1V group had less Advax-SM, the nanoparticle Q@NP-adjuvanted S-protein percentage of adverse events compared with BNT162b2 vaccine, and the TLR7/8 agonist-adjuvanted recombinant group, most of mild intensity. spike protein vaccine with published data in animal models, while the NVX-CoV2373, Spikogen®, S-26019-b, V-01, Implications of all the available evidence Nanocovax, RDB Vaccine (Abdala), CoV2 preS dTM-AS03 and PHH-1V can elicit protection against current circulating SARS- MVC-COV1901 had publications for their registered clinical CoV-2 variants of concern. This protein-based vaccine trials. In addition, we searched public databases, as the represents a safe, well tolerated, and efficient option for “global database of COVID-19 vaccinations” (https:// heterologous booster, improving the range of vaccines ourworldindata.org/covid-vaccinations) to have an update in available to cover diverse global demands and reach those COVID-19 vaccines doses administered, and the “COVID-19 with none or low vaccine coverage. Vaccine Development and Approvals Tracker, 2020” from day 14 post-vaccination, asserting its value as a Regarding the inactivated virus vaccines, the BBIBP- basis for immunological memory and its role in the CorV (Sinopharm), Coronavac (Sinovac) and Covaxin rapid response to pathogen re-exposure, conferring (Bharat Biotech) have been approved by the World protection from severe disease even with lower levels of Health Organization (WHO) for emergency use. On the antibodies.4–7 other hand, the protein-based vaccines contain the spike The types of vaccines developed for SARS-CoV-2 protein in different forms and combinations with ad- carry the virus’s genetic material, contain inactivated juvants. The NVX-CoV2373 (Novavax) corresponds to a virus or are protein-based. The first type contains the subunit vaccine, containing the full-length spike protein genetic information for the biosynthesis of the spike with the Matrix M1 adjuvant, also approved by the WHO glycoprotein, such as mRNA (BNT162b2 [Pfizer/Bio- and EMA. The CoV2 preS dTM adjuvanted vaccine NTech] and mRNA-1273 [Moderna]) and adenoviral (Sanofi/GSK) is a prefusion spike recombinant protein vector vaccines (Ad26.COV2.S [Janssen] and ChAdOx vaccine recently approved by the EMA. In contrast to [Oxford-AstraZeneca]).8 Vaccines based on the mRNA mRNA, protein-based vaccines can be stored under technology have inherent limitations: the liability of the refrigerated conditions.6,10,11 mRNA molecule requires that mRNA-based vaccines Although various SARS-CoV-2 vaccines have been are encapsulated in lipid nanoparticles and kept at very approved, there is still the utmost need for options low temperatures (between −80 ◦ C and −20 ◦ C).9 able to meet global demands of immunizations.12 www.thelancet.com Vol 28 May, 2023 3 Articles Vaccination campaigns protect against severe disease, Eligibility criteria were individuals aged 18 years or reduce viral load and therefore lower transmission older, who had received two doses of the BNT162b2 risk, which implies a public health benefit.13 Booster vaccine at least 182 days and less than 365 days after vaccination has become necessary because immunity their second dose; Body Mass Index (BMI) between 18 diminishes with time and new variants emerge. and 40 kg/m2; negative SARS-CoV-2 PCR test at the Increasing evidence shows that heterologous vaccina- time of enrolment; willingness to avoid all other vac- tion is safe, with similar reactogenicity and adverse cines within 4 weeks before and after vaccination in this events compared to a homologous schedules.14–17 study (seasonal influenza vaccination was allowed if it Several recent clinical studies indicate that heterolo- was received at least 14 days before or after the study gous schedules may, in fact, elicit a more robust im- booster). Key exclusion criteria included pregnancy or mune response and higher antibody titres, with breastfeeding; an ongoing serious psychiatric condition; consequent higher effectiveness, supporting a mix- history of respiratory disease requiring daily medica- and-match approach.18–20 These data suggest that tions; history of significant cardiovascular disease; his- there is a strong case for more vaccine platforms to tory of neurological or neurodevelopmental conditions; broaden vaccination options worldwide. ongoing malignancy or recent diagnosis of malignancy Recombinant protein-based vaccines possess a good in the last five years excluding basal cell and squamous safety profile, no risk of genome integration, no live cell carcinoma of the skin; any confirmed or suspected components, and are suitable for people with compro- autoimmune, immunosuppressive or immunodefi- mised immune systems, showing high productivity ciency disease/condition (iatrogenic or congenital) yields and good stability profiles.21,22 PHH-1V is a including human immunodeficiency virus infection; protein-based vaccine intended for the prevention of use of immunosuppressants; coagulation or bleeding COVID-19 caused by SARS-CoV-2. This vaccine is based disorder; chronic liver disease; history of SARS-CoV-2 on a fusion heterodimer consisting of the spike RBD infection; close contact with anyone positive for SARS- sequence from the SARS-CoV-2 Beta (B.1.351) and CoV-2 infection within 15 days before screening and Alpha (B.1.1.7) variants, formulated with an oil-in-water life expectancy of less than 12 months. Full inclusion emulsion based on squalene (SQBA) intended for and exclusion criteria are provided in the study protocol intramuscular administration.23–25 PHH-1V comprises (Supplementary data). three key mutations of high relevance for previously and The trial was conducted in accordance with the currently circulating SARS-CoV-2 variants: K417N, Declaration of Helsinki, the Good Clinical Practice E484K and N501Y.26 Previous studies revealed that the guidelines, and national regulations. The study protocol E484K and N501Y mutations are present in the RBD of was reviewed and approved by the Spanish Agency of the spike protein of Beta, Gamma, and Mu variants; and Medicines and Medical Devices (AEMPS) as well as both the K417N and N501Y mutations present in the Independent Ethics Committee from the Hospital PHH-1V antigen are common to the Beta and to the Clínic de Barcelona. Written informed consent was ob- currently predominant Omicron variants.27 tained from all participants before enrolment. The bio- A first-in-human Phase I/IIa study dose-escalation, logic biosafety committee of the Research Institute randomized, double-blinded, active-comparator contro- Germans Trias i Pujol approved the execution of SARS- lled clinical trial in 30 healthy adults demonstrated that CoV-2 experiments at the BSL3 laboratory of the Centre the PHH-1V vaccine is safe and well-tolerated in healthy for Bioimaging and Comparative Medicine (CSB-20- young adults, with even fewer reported solicited adverse 015-M8). events compared to the control, BNT162b2.28 The HH-2 study is an ongoing multicentre, rando- Randomisation and masking mised, active-controlled, double-blind, non-inferiority This study was double-blinded; participants, site staff, Phase IIb trial that aims to assess the immunogenicity including clinical staff involved in study drug prepara- and safety of PHH-1V as a heterologous boost versus a tion and administration, laboratory analysis staff, the homologous boost in individuals after receiving a pri- sponsor, and the Clinical Research Organisation (CRO) mary vaccination series of BNT162b2. Here we show were blinded to treatment assignment/allocation. Un- interim data up to 98 days after administration of the blinded hospital pharmacists or other qualified booster for the Phase IIb HH-2 study. personnel prepared the booster dose, and unblinded site staff members, who were not otherwise involved with Methods study procedures (except for blood extraction), admin- Study design and participants istered treatment to subjects. This multicentre, randomised, active-controlled, double- Subjects were randomly assigned in a 2:1 ratio to blind, non-inferiority Phase IIb trial where immunoge- receive a booster dose of vaccine (third immunization) nicity and safety of the PHH-1V vaccine were assessed, either with the HIPRA PHH-1V vaccine (PHH-1V was carried out in 10 centres in Spain (appendix group) or with the Pfizer/BioNTech BNT162b2 vaccine pp.10–11 Supplementary material). (BNT162b2 group). Subjects were allocated to treatment 4 www.thelancet.com Vol 28 May, 2023 Articles using an Interactive Response Technology (IRT). The 14, 28 and 98 days after boosting using the Elecsys Anti- allocation sequence was stratified by age group: SARS-CoV-2 S immunoassay (Roche Diagnostics). The approximately 90% of the total recruited participants titre of neutralizing antibodies was also analysed in a were in the 18–64 years group, and 10% in the group 65 subset of subjects, corresponding to approximately 20% years or older. This article refers to the data analysis of the total subjects included in the study, representing obtained on days 14, 28 and 98. the first 58 participants vaccinated during this HH-2 trial (for BNT162b2 active control, n = 24; for PHH-1, Study vaccine n = 34) by infectious SARS-CoV-2 neutralisation test The PHH-1V vaccine was provided in a vial containing (VNA) and measured as ID50. The geometric mean titre one dose of 0.5 mL (40 μg), ready to use, and stored at (GMT) and the geometrical mean fold rise (GMFR) were 2–8 ◦ C. The adjuvant is an oil-in-water emulsion based calculated for each parameter. on squalene produced by HIPRA (SQBA). The safety assessment included the incidence of Comirnaty was supplied as a frozen suspension in a solicited local and systemic reactions, unsolicited local multidose vial that must be thawed and diluted before and systemic adverse events reported in the patient’s use. One vial (0.45 mL) contained 6 doses of 0.3 mL diary. Additionally, the incidence of SARS-CoV-2 in- after dilution. Comirnaty was stored in an ultra-low fections, including severe infections, the requirement of temperature freezer between −90 and −60 ◦ C, until hospital and intensive care unit (ICU) admissions, and the expiry date printed on the label. Alternatively, deaths associated with COVID-19, if any, through the Comirnaty could be stored at −25 to −15 ◦ C for up to 2 end of the study will be assessed. In the safety analysis, weeks. Vials were kept frozen and protected from light, adverse events were coded using the MedDRA Version in the original package, until ready to use. 24.1 coding system. A label was used to mask the syringe to visually The T-cell mediated immune response against the distinguish the two treatments. The blind was broken as SARS-CoV-2 spike glycoprotein was assessed after the soon as the last study participant reached day 28 after in vitro peptide stimulation of peripheral blood mono- the study booster dose. Participants would then know nuclear cells (PBMC) followed by interferon gamma which vaccine they had received and were able to decide enzyme-linked immune absorbent spot (IFN-γ ELI- whether they wanted to receive, or not, a commercial Spot) and intracellular cytokine staining (ICS) (details COVID 19-vaccine. In the event a participant decided to of procedures are described in the appendix (pp.3 receive a commercial COVID-19 vaccine booster ac- Supplementary material). cording to the vaccination schedule established by local authorities, an extra visit was scheduled, with the Outcomes recommendation to wait 3 months after administration The primary endpoints were humoral immunogenicity of HIPRA’s vaccine to receive the commercial booster. measured by changes in levels of neutralizing anti- bodies (PBNA) against the ancestral Wuhan-Hu-1 strain Procedures after the PHH-1V or the BNT162b2 boost, and the safety The study visits were scheduled on day 0, day 14, day 28, and tolerability of PHH-1V as a boost. The secondary day 98, day 182 and day 364. This article reports data up endpoints were to compare changes in levels of neu- to day 98. Participants who, at that moment, met eligi- tralizing antibodies against different variants of SARS- bility criteria were vaccinated in the visit on day 0. The CoV-2, the T-cell responses towards the SARS-CoV-2 BNT162b2 vaccine was given as a 0.3 mL (30 μg) and the spike glycoprotein peptides and to compare the PHH-1V as a 0.5 mL (40 μg) by intramuscular injection changes in immunogenicity measured by wild type into the deltoid muscle. The first 30 participants were SARS-CoV-2 neutralisation test (VNA) only in a subset observed for 60 min and monitored during the of subjects, 20% approximately of the total subjects following 72 h by phone. Other participants were included in the study. The exploratory endpoint was to observed for 30 min and contacted again on day 7. assess the number of subjects with SARS-CoV-2 in- During the day-0 visit, participants were given a hard fections ≥14 days after PHH-1V booster. copy diary to record local and systemic solicited re- actions within the 7 days after vaccination. Statistical analysis The neutralization titres of antibodies were deter- The sample size was calculated in accordance with the mined by the inhibitory dilution 50 (ID50, reported as Food and Drug Administration (FDA) Guidance for the half maximal inhibitory dilution, which is the Industry on Clinical Data Needed to Support the reciprocal dilution of the test serum required to inhibit Licensure of Seasonal Inactivated Influenza Vaccines. the virus effect by 50%) using a pseudovirion-based Noninferiority for a new influenza vaccine product neutralisation assay (PBNA) as described previously.29 could be claimed if the upper bound of the two-sided The immunogenicity against the SARS-CoV-2 spike 95% confidence interval (CI) surrounding the ratio of glycoprotein was assessed by the percentage of subjects GMT for the control to investigational product does not having a ≥4-fold increase in the binding antibodies titre exceed 1.5. Superiority is concluded if the upper bound www.thelancet.com Vol 28 May, 2023 5 Articles of the 95% confidence interval of the ratio of GMTs margin (NIm); equivalently, the difference in log (BNT162b2: PHH-1V) is below 1. Given uncertainty in (GMT) exceeds log (NIm); ii) Alternative hypothesis, H1: the immune response and variability, the Phase IIb part the ratio of GMTs is below NIm; equivalently the dif- of this study was planned with a reduced non-inferiority ference in log (GMT) is less than log (NIm). The NIm margin of 1.4 to ensure sufficient sample size for safety established for this study was 1.4, whereby the upper and immunogenicity assessments. bound of the 95% confidence interval (95% CI) had to Considering these assumptions, and with a 2:1 ran- be lower than this value to accept the null hypothesis domisation ratio, group sample sizes of 301 and 151, and was defined for each endpoint separately. respectively, would achieve 90% power to detect non- For the immunogenicity analysis, concerning the inferiority using a one-sided 2.5% significance level, log10-transformed PBNA and VNA data on neutralising two-sample t-test using a SDlog = 0.45 for both treat- antibody titres against the SARS-CoV-2 variants and ments. Assuming a 25% withdrawal rate, a total of 602 Wuhan-Hu-1 strain as well as the total binding anti- subjects (401 in PHH-1V group, 201 in the BNT162b2 bodies data, linear mixed effects models were employed. active control group) were previewed to be randomised in In these models, the treatment group, the age group, the this study. When reaching the 602 participants, the visit (baseline and day 14 for VNA and day 28 and 98 for hospitals already had additional appointments for vacci- PBNA) and the treatment-by-visit interaction were nation of participants, and in agreement with HIPRA included as fixed effects, while the site and the subject- and the CRO, it was decided to allow the hospitals to nested-to-site factors were included as random effects in complete the vaccinations already scheduled during one random intercept models. In all models, a compound additional week approximately. Otherwise, the hospitals symmetry variance-covariance matrix structure was would have to cancel the vaccinations already scheduled. used, and restricted maximum likelihood was employed A total of 782 subjects were enrolled into the Phase IIb for parameter estimation. Denominator degrees-of- part of the study, stratified 2:1 in the following age freedom were calculated using the Kenward-Roger’s groups: 18–64 years and equal or over 65 years. approximation. Validity of the models was assessed The following analyses of populations were included graphically (i.e., quantile–quantile plots, residuals plots). in this study: Intention-to-treat (ITT) population, For pairwise and interaction contrasts, estimated mar- including all subjects who were randomly assigned to ginal means were calculated and compared. treatment, regardless of the subject’s treatment status in Similarly, linear mixed effects models were the study; modified intention-to-treat (mITT) popula- employed for the analysis of the T-cell response data tion, consisting of all subjects in the ITT who met the (ELISpot and ICS). Importantly, the proportions of screening criteria and received one vaccine dose (sub- positive cells were transformed using the angular jects who tested positive for COVID-19 within 14 days of (arcsine-square root) transformation. In these models, the receiving study drug were excluded); mITT3 (98) the visit, the treatment, the peptide pool stimulus, and population, including all subjects in the mITT without the three-way interaction were included as fixed effects, COVID-19 infections recorded via adverse event while the subject identifier was included as a random reporting prior to their Day 98 visit date; immunoge- effect. In all models, a heterogeneous first order autor- nicity (IGP) population, including all subjects in the egressive variance-covariance matrix structured was mITT who had a valid immunogenicity test result before employed, and restricted maximum likelihood was used receiving study drug and at least one valid result after for parameter estimation. Validity of the models was dosing; per-protocol (PP), all subjects in the mITT who assessed graphically (i.e., quantile–quantile plots, re- received one vaccine dose and had no major protocol siduals plots). For pairwise and interaction contrasts, deviations, and documented by the Sponsor prior to data estimated marginal means were calculated and base lock and unblinding that impact critical or key compared. To analyse the antibody neutralisation titre study data; safety (SP), comprised of all randomised against SARS-CoV-2 (Wuhan-Hu-1, Beta, Delta, and subjects who received one vaccine dose, and were ana- Omicron BA.1) as measured by ID50 by PBNA or VNA, lysed according to the treatment received. The humoral a mixed effects model for repeated measures (MMRM) immunogenicity analysis was done in both mITT and was carried out on log transformed data. Summary mITT3 (98) populations, and the T-cell response from a statistics for the log10 transformations for each individ- subset of the total subjects included in three of the study ual sample were calculated based on the log10-trans- sites. The safety analysis was conducted in the safety formed titres at baseline, on day 14, day 28 and day 98, population and the assessment of COVID-19 cases in and are presented for the mITT population. the PP population. The GMFR analysis was performed for participants The humoral immunogenicity analyses tested the who, after a booster dose, had a ≥4-fold change in following hypotheses to show non-inferiority of PHH- binding antibodies titre from baseline to day 14, day 28 1V when compared to the BNT162b2 booster vaccine: and day 98 (responders), summarised by the number i) Null hypothesis, H0: the ratio of the GMTs and proportion of responders along with an exact 95% (BNT162b2: PHH-1V) exceeds the non-inferiority Clopper-Pearson CI for the proportion. 6 www.thelancet.com Vol 28 May, 2023 Articles For the cellular immunogenicity analysis, a MMRM PHH-1V (n = 522) or BNT162b2 (n = 260) booster of T-cell data (IFN-γ Spot-Forming Units [SFU]/106 vaccine groups (Fig. 1). A total of 504 participants in the PBMCs) was employed, using the angular-transformed PHH-1V group and 247 participants in the BNT162b2 proportions as the response variable. group were included in the mITT population. For the SP For the incidence of COVID-19 data are shown as the population, 513 and 252 participants were included in number of events and percentage of participants the PHH-1V and BNT162b2 groups, respectively. Five affected in the PP population. An exact 95% Clopper- hundred and three subjects were included in the PHH- Pearson CI for the proportion of each endpoint was 1V group and 246 participants in the BNT162b2 group also presented. (IGP population). Since this trial is ongoing, the PP This trial is registered at ClinicalTrials.gov, population has not yet been defined as protocol de- NCT05142553. viations are still being assessed. Baseline characteristics were well balanced between the two groups, where the Role of the funding source median age of participants was 42 years (range 19–76 This study was sponsored by HIPRA SCIENTIFIC, years; p = 0.65), 484 (63.3%) were female (p = 1.00) and S.L.U (HIPRA). HIPRA was involved in the study 755 (98.7%) identified as White (p = 1.00; Table 1). design; in the collection, analysis, and interpretation of The GMT and GMFR of neutralizing antibodies data; in writing of the report; and in the decision to determined by PBNA on day 14, 28 and 98 post-booster submit the paper for publication. are shown in Table 2 and Fig. 2A for the mITT popula- tion. For the primary endpoint of humoral immunoge- nicity measured by changes in levels of neutralizing Results antibodies (PBNA) against the ancestral Wuhan-Hu-1 From 15 November 2021, 862 participants were scr- strain, the GMT on day 14 was 1965.79 (95% CI: eened, of whom, 782 adults were randomly assigned to [1712.26, 2256.85]) for the PHH-1V group and 3309.29 862 participants assessed for eligibility 80 ineligible 72 did not meet eligibility criteria 8 withdrawal 782 enrolled 782 randomised 522 assigned to PHH-1V 260 assigned to BNT162b2 booster intention-to-treat booster intention-to-treat population population 9 did not 8 did not receive study’s receive study’s drug drug 513 included in safety analysis 252 included in safety analysis 5 not meeting 9 not meeting inclusion/exclusion inclusion/exclusion criteria after criteria after vaccination vaccination 504 included in modified 247 included in modified intention-to-treat and per- intention-to-treat and per- protocol population protocol population 1 not valid 1 not valid immunogenicity immunogenicity test/ test/ valid result valid result 503 included in the 246 included in the immunogenicity population immunogenicity population Fig. 1: Trial Profile. PHH-1V = PHH-1V vaccine, HIPRA. BNT162b2 = BNT162b2 vaccine, Pfizer–BioNTech. www.thelancet.com Vol 28 May, 2023 7 Articles PHH-1V BNT162b2 Total Confidence Interval p value Number enrolled 513 (100.0) 252 (100.0) 765 (100.0) Age, years All age groups, median (range) 42.0 (19.0–76.0) 40.0 (20.0–74.0) 42.0 (19.0–76.0) 0.50 [−1.72, 2.72]† 0.65 18–65 475 (92.6) 234 (92.9) 706 (92.7) ≥65 38 (7.4) 18 (7.1) 56 (7.3) 0.96 [0.53, 1.73]‡ 1.00 Sex Female 325 (63.4) 159 (63.1) 484 (63.3) Male 188 (36.6) 93 (36.9) 281 (36.7) 0.99 [0.72, 1.37]‡ 1.00 Race White 505 (98.4) 250 (99.2) 755 (98.7) Asian 4 (0.8) 1 (0.4) 5 (0.6) 1.00 American Indian or Alaska native 2 (0.4) 0 (0.0) 2 (0.3) Other 2 (0.4) 1 (0.4) 3 (0.4) Ethnicity Hispanic or Latino 170 (33.1) 77 (30.6) 247 (32.3) 1.13 [0.8, 1.56]‡ 0.51 Not Hispanic or Latino 343 (66.9) 175 (69.4) 518 (67.7) BMI, median (range) 24.03 (16.98–39.78) 23.73 (17.90–39.86) 23.92 (16.98–39.86) 0.16 [−0.47, 0.79]† 0.61 Time between second dose and booster, na 446 218 664 – – Days, median (range) 296.5 (162, 325) 296.0 (182, 318) 296.0 (162, 325) Data are n (%) unless otherwise specified. Confidence intervals are shown as † mean difference, 95% CI for the comparison of basal continuous variables or ‡ Odds Ratio, 95% CI for the comparison of dichotomous variables between groups. aParticipants who did not have the complete date in the second or third dose and those in which the subtraction “3rd dose -2nd dose” was negative were excluded from the analysis. BMI = body-mass index. Table 1: Baseline characteristics of participants included in the analysis. (95% CI [2811.97, 3894.56]) for the BNT162b2 group, and 63.5% for the BNT162b2 group) (Supplementary with a GMT ratio of 1.68 (95% CI [1.44, 1.97]; p < 0.0001), Table S2). The most frequent post-vaccination solicited which could not demonstrate non-inferiority of the systemic adverse events, for the PHH-1V and the PHH-1V vaccine to the BNT162b2 vaccine in immune BNT162b2 groups on day 1, respectively, were fatigue response. The same was observed for day 28, with a GMT (16.0% and 35.3%), headache (14.2% and 27.8%), ratio of 1.31 (95% CI [1.12, 1.54]; p = 0.0007). However, muscle pain (11.7% and 29.4%), and fever (0.6% and on day 98, the GMT ratio was 0.86 (95% CI [0.60, 1.22]; 7.1%) (Supplementary Table S2). p = 0.40), which shows non-inferiority of the PHH-1V Overall, 458 (89.3%) subjects in the PHH-1V group vaccine to the BNT162b2 vaccine immune response. and 238 (94.4%) subjects in the BNT162b2 group expe- The GMFR data in Table 2 confirm the observed for rienced at least one adverse event (p = 0.0219) through GMT. A booster vaccination with PHH-1V could not day 28 after administration. Specifically, events of mild demonstrate non-inferiority for Wuhan-Hu-1 on day 14 intensity were reported by 66.7% and 57.9% of subjects (ratio 1.74, 95% CI [1.44, 2.09]; p < 0.0001) and day 28 in the PHH-1V and BNT162b2 groups, respectively (ratio 1.35, 95% CI [1.13, 1.63]; p = 0.0013), but it did so (p = 0.02; Fig. 3 and Supplementary Table S3). The most for day 98 (ratio 0.89, 95% CI [0.61, 1.28]; p = 0.52). These frequent adverse events were injection site pain (79.7% humoral immunogenicity results were additionally and 89.3%; p = 0.0010), headache (31.2 and 40.1%; confirmed in the mITT3 (98) population (excluding those p = 0.0190) and fatigue (27.5% and 42.1%; p = 0.001) for subjects who reported COVID-19 infections trough day the PHH-1V and the BNT162b2 groups, respectively 98; Supplementary Table S1), where non-inferiority was (Fig. 3 and Supplementary Table S3). Treatment-related demonstrated in day 98, with a GMT and a GMFR ratio adverse events were reported in 434 subjects (1384 ev- of 0.84 [0.59, 1.20] (p = 0.34) and 0.92 [0.63, 1.34] ents, 84.6%) in the PHH-1V group and 231 subjects (975 (p = 0.65), respectively. events, 91.7%) in the BNT162b2 group (p = 0.0061). No Concerning the primary endpoint of safety and deaths were reported in the study in either of the PHH- tolerability, complete data for the vaccination diary are 1V or BNT162b2 group and one SAE was reported for the shown in the appendix up to day 7 post-boost PHH-1V group (0.2%, p = 1.0; Supplementary Table S3). (Supplementary Table S2). Briefly, the most frequent For the secondary endpoint of humoral immunoge- solicited local reactions on day 1 were pain (51.1% and nicity against different variants of SARS-CoV-2, the 69.8% for the PHH-1V and the BNT162b2 group, GMT for the beta variant on day 14 was 4335.87 (95% CI respectively) and tenderness (48.5% for the PHH-1V [3760.91, 4998.72]) in the PHH-1V group and 2672.16 8 www.thelancet.com Vol 28 May, 2023 www.thelancet.com Vol 28 May, 2023 Variant PHH-1V (n = 504) BNT162b2 (n = 247) Day 14 Day 28 Day 98 Day 14 Day 28 Day 98 Neutralizing antibodies Wuhan-Hu-1 n (%) 500 (99.2) 496 (98.4) 78 (15.5) 241 (97.6) 244 (98.8) 42 (17.0) GMT 1965.79 [1712.26, 2256.85] 2244.22 [1954.53, 2576.85] 1119.74 [886.67, 1414.07] 3309.29 [2811.97, 3894.56] 2944.60 [2502.10, 3464.12] 960.29 [708.77, 1301.07] GMT ratio 1.68 [1.44, 1.97]; p < 0.0001 1.31 [1.12, 1.54]; p = 0.0007 0.86 [0.60, 1.22]; p = 0.40 GMFR 22.79 [19.83, 26.18] 26.01 [22.63, 29.90] 12.98 [9.73, 17.32] 39.60 [32.44, 48.33] 35.23 [28.88, 42.98] 11.49 [7.76, 17.02] GMFR ratio 1.74 [1.44, 2.09]; p < 0.0001 1.35 [1.13, 1.63]; p = 0.0013 0.89 [0.61, 1.28]; p = 0.52 Beta n (%) 500 (99.2) 496 (98.4) 78 (15.5) 241 (97.6) 244 (98.8) 42 (17.0) GMT 4335.87 [3760.91, 4998.72] 3813.26 [3307.12, 4396.86] 1915.50 [1500.75, 2444.87] 2672.16 [2247.87, 3176.54] 2480.27 [2087.63, 2946.75] 1075.00 [780.00, 1481.59] GMT ratio 0.62 [0.52, 0.74]; p < 0.0001 0.65 [0.54, 0.78]; p < 0.0001 0.56 [0.38, 0.82]; p = 0.003 GMFR 64.90 [56.18, 74.97] 57.07 [49.39, 65.96] 28.67 [21.19, 38.78] 43.84 [35.64, 53.94] 40.69 [33.11, 55.02] 17.64 [11.68, 26.64] GMFR ratio 0.68 [0.56, 0.82]; p = 0.0001 0.71 [0.59, 0.86]; p = 0.0006 0.62 [0.42, 0.91]; p = 0.0142 Delta n (%) 500 (99.2) 496 (98.4) 78 (15.5) 241 (97.6) 244 (98.8) 42 (17.0) GMT 1475.09 [1280.69, 1698.99] 1719.68 [1492.87, 1980.96] 1796.11 [1423.79, 2265.78] 1486.52 [1258.98, 1755.18] 1512.30 [1281.47, 1784.72] 938.77 [694.28, 1269.36] GMT ratio 1.01 [0.86, 1.18]; p = 0.92 0.88 [0.75, 1.03]; p = 0.11 0.52 [0.37, 0.74]; p = 0.0003 GMFR 32.72 [28.57, 37.46] 38.14 [33.29, 43.69] 39.84 [30.02, 52.86] 35.92 [29.57, 43.64] 36.55 [30.11, 44.37] 22.69 [15.43, 33.38] GMFR ratio 1.10 [0.92, 1.32]; p = 0.31 0.96 [0.80, 1.15]; p = 0.64 0.57 [0.40, 0.82]; p = 0.0024 Omicron BA.1 n (%) 500 (99.2) 496 (98.4) 78 (15.5) 241 (97.6) 244 (98.8) 42 (17.0) GMT 2036.14 [1752.22, 2366.06] 1512.00 [1301.00, 1757.21] 617.70 [483.04, 789.90] 1210.73 [1014.49, 1444.94] 992.07 [831.71,1183.34] 349.18 [253.68, 480.63] GMT ratio 0.59 [0.50, 0.71]; p < 0.0001 0.66 [0.55, 0.78]; p < 0.0001 0.57 [0.39, 0.82]; p = 0.0028 GMFR 62.71 [54.35, 72.34] 46.56 [40.35, 53.74] 19.02 [14.11, 25.64] 41.97 [34.18, 51.54] 34.39 [28.03, 42.19] 12.10 [8.05, 18.19] GMFR ratio 0.67 [0.55, 0.81]; p < 0.0001 0.74 [0.61, 0.89]; p = 0.0018 0.64 [0.43, 0.93]; p = 0.0207 Binding antibodies GMFR Fold Change (≥4-fold) n (%) 490 (98.4) 484 (97.6) 78 (15.5) 238 (98.8) 240 (98.4) 42 (17.0) Odds 61.85 [30.69, 124.65] 40.70 [22.96, 72.16] 22.05 [8.90, 54.62] 80.83 [25.75, 253.76] 60.34 [22.46, 162.06] 20.87 [6.47, 67.33] Odds ratio 1.31 [0.34, 5.00]; p = 0.70 1.48 [0.47, 4.64]; p = 0.50 0.95 [0.22, 4.16]; p = 0.94 PHH-1V (n = 504) BNT162b2 (n = 247) Day 14 Day 14 SARS-CoV-2 neutralisation test (VNA)—Omicron BA.1 Adjusted treatment mean 3.38 [3.04, 3.71] 3.02 [2.72, 3.33] GMT 2379.71 [1109.68, 5103.31] 1055.61 [526.33, 2117.11] GMT ratio 0.44 [0.30, 0.65]; p = 0.0001 CI = confidence interval; GMT = Geometric Mean Titre; GMFR = Geometric Mean Fold Rise; LS mean = least square mean; SE= Standard Error. n (%), refers to subjects with data; GMT is shown as adjusted treatment mean [95% CI]; GMT ratio is Articles shown as BNT162b2 active control vs PHH-1V [95% CI] followed by p-value for ratio = 1. GMFR is shown as fold rise of adjusted treatment means between timepoints [95% CI]; GMFR ratio is shown as BNT162b2 active control vs PHH-1V [95% CI] followed by p-value for ratio = 1; GMFR fold change is shown for subjects with ≥4-fold change in binding antibodies; odds are shown as back-transformed adjusted treatment LS means [95% CI]; Treatment effect is shown for “BNT162b2 vs PHH- 1V” as the odds ratio [95% CI] followed by p-value for odds ratio = 1. For SARS-CoV-2 neutralisation test (VNA), Adjusted treatment mean is shown for BNT162b2 active control vs PHH-1V as LS mean [95% CI]. GMT is shown for adjusted treatment mean [95% CI]. GMT ratio is the GMT for treatment ratio for BNT162b2 active control vs PHH-1V [95% CI] followed by p-value for ratio = 1. Table 2: Analysis of neutralizing and binding antibodies against SARS-CoV-2 variants on days 14, 28 and 98 post-vaccination boost. 9 Articles a Wuhan−Hu−1 Beta Vaccine ns *** *** ns ns *** *** ** PHH−1V 5 5 BNT162b2 4 4 log10IC50 log10ID50 3 3 2 2 1 1 Baseline D14 D28 D98 Baseline D14 D28 D98 Delta Omicron ns ns ns *** ns *** *** ** 5 5 4 4 log10IC50 log10ID50 3 3 2 2 1 1 Baseline D14 D28 D98 Baseline D14 D28 D98 b ns *** *** *** Vaccine PHH−1V 5 BNT162b2 Total binding antibody levels ( log10 U/ml) 4 3 2 1 Baseline D14 D28 D98 c Omicron BA.1 4.5 Vaccine * Neutralizing anti-SARS-CoV-2 PHH−1V 4 BNT162b2 (VNA) (log10ID50) 3.5 3 2.5 2 1.5 Group 10 www.thelancet.com Vol 28 May, 2023 Articles Total Adverse events Injection site pain Headache Fatigue Myalgia Injection site duration Injection site erythema 100% * ** 75% Percentage (%) 50% * *** 25% *** ** *** 0% PHH-1V BNT162b2 PHH-1V BNT162b2 PHH-1V BNT162b2 PHH-1V BNT162b2 PHH-1V BNT162b2 PHH-1V BNT162b2 PHH-1V BNT162b2 Severe Moderate Mild Fig. 3: Adverse Events by MedDRA Preferred Term (PT) by treatment group through day 28 after booster administration reported in ≥10.0% of Overall Subjects. Data are shown as the percentage of subjects in relation to the safety population (PHH-1V, N = 513 and BNT162b2, N = 252), indicating the intensity of adverse events (mild, moderate and severe). If a subject experienced more than one adverse event, the subject is counted once for each PT choosing the most severe. PTs are ordered in decreasing frequency of the total number of subjects with each adverse event in PHH-1V group. *p < 0.05; **p < 0.001; ***p < 0.0001 (data from Supplementary Table S3). (95% CI [2247.87, 3176.54]) in the BNT162b2 group, mITT3 (98) population (Supplementary Table S1), with a GMT ratio of 0.62 (95% CI [0.52, 0.74]; except for the inconclusive result for the Delta variant in p < 0.0001), indicating a superiority of the PHH-1V day 28, which demonstrated a non-inferiority when vaccine. The GMT ratios of 0.65 (95% CI [0.54, 0.78]; compared to BNT162b2. p < 0.001) and 0.56 (95% CI [0.38, 0.82]; p = 0.003) on Concerning the GMFR ratios of the different days 28 and 98, respectively, confirm the superiority of assessed variants of SARS-CoV-2, the data in Table 2 the PHH-1V for the Beta variant. Regarding the Delta confirm the results observed for GMT on day 14. A variant, the GMT on day 14 was 1475.09 (95% CI booster vaccination with PHH-1V provided non- [1280.69, 1698.99]) and 1486.52 (95% CI [1258.98, inferiority for the Delta variant (ratio 1.10, 95% CI 1755.18]) for the PHH-1V and BNT162b2 groups, [0.92, 1.32]; p = 0.31), and elicited statistically signifi- respectively, with a GMT ratio of 1.01 (95% CI [0.86, cant higher levels of neutralizing antibodies for the 1.18]; p = 0.92), demonstrating non-inferiority of the Beta (ratio 0.68, 95% CI [0.56, 0.82]; p = 0.0001) and PHH-1V vaccine to BNT162b2. On days 28 and 98, the Omicron BA.1 (ratio 0.67, 95% CI [0.55–0.81]; GMT ratio of 0.88 (95% CI [0.75, 1.03]; p = 0.11) and p < 0.0001) variants, thus, demonstrating superiority of 0.52 (95% CI [0.37, 0.74]; p = 0.0003) indicates an the PHH-1V booster against Beta and Omicron BA.1 inconclusive result and a superiority of the PHH-1V variants. On day 28, a booster vaccination with PHH- vaccine for the Delta variant, respectively. For the Om- 1V demonstrates non-inferiority for Delta variant (ra- icron BA.1 variant, the PHH-1V and the BNT162b2 tio 0.96, 95% CI [0.80, 1.15]; p = 0.64), and a superiority boosted groups presented a GMT of 2036.14 (95% CI of PHH-1V was observed for Beta and Omicron BA.1 [1752.22, 2366.06]) and 1210.73 (95% CI [1014.49, variants (ratios of 0.71, 95% CI [0.59, 0.86]; p = 0.0006 1444.94]), respectively, and a GMT ratio of 0.59 (95% CI and 0.74, 95% CI [0.61, 0.89]; p = 0.0018, respectively). [0.50, 0.71]; p < 0.0001) on day 14, thus demonstrating On day 98 post-boost, the GMFR ratio for the Beta superiority of the PHH-1V vaccine to the BNT162b2 (ratio 0.62, 95% CI [0.42, 0.91]; p = 0.0142), Delta (ratio vaccine against the Omicron BA.1 SARS-CoV-2 variant. 0.57, 95% CI [0.40, 0.82]; p = 0.0024), and Omicron This result is further confirmed on days 28 and 98, with BA.1 (ratio 0.64, 95% CI [0.43, 0.93]; p = 0.0207) vari- a GMT ratio of 0.66 (95% CI [0.55, 0.78]; p < 0.001) and ants demonstrates superiority of the PHH-1V vaccine 0.57 (95% CI [0.39, 0.82]; p = 0.0028), respectively. All to the BNT162b2 vaccine against these variants. These results for GMT and GMT ratios were confirmed in the results were further confirmed when the data analysis Fig. 2: Analysis of the antibody response to PHH-1V vaccination. (A) Neutralizing antibody responses against multiple SARS-CoV-2 variants by PBNA in the mITT population. Comparison of neutralizing antibody titres at baseline, day 14, day 28 and day 98, between the PHH-1V and BNT162b2-vaccinated groups. (B) SARS-CoV-2 binding antibodies titre. Comparison of total binding antibody titres at baseline, day 14, day 28 and day 98, between the PHH-1V and BNT162b2-vaccinated groups. The analysis has been performed considering only responder subjects (≥4 fold-change in binding antibodies titre from baseline to day 14, day 28 or day 98). (C) SARS-CoV-2 neutralisation assay (VNA) for the Omicron variant (B.1.1.529). VNA assay was performed with serial dilutions of heat-inactivated serum samples from individuals receiving a PHH-1V or BNT162b2 boosting vaccine on day 14. VNA titres are plotted as ID50 (the reciprocal dilution inhibiting 50% of the cytopathic effect). *p-value for ratio = 1, BNT162b2 active control vs PHH-1V; p = 0.0001. Adjusted treatment mean is shown for VNA analysis for BNT162b2 active control (n = 24) vs PHH-1V (n = 34) as LS mean (95% CI). GMT is shown for adjusted treatment mean (95% CI). GMT ratio is the GMT for treatment ratio for BNT162b2 active control vs PHH-1V (95% CI) followed by p-value for ratio = 1. Boxplots with grey dots and blue dots refer to PHH-1V and BNT162b2, respectively. The red dashed lines shown in the boxplots 2A and 2B indicate the non-inferiority limit, calculated as the GMT value of BNT162b2 divided by 1.4. Statistically significant differences are shown as * for p ≤ 0.05; ** for p ≤ 0.01; *** for p ≤ 0.001. Non- significant comparisons have been indicated with “ns”. www.thelancet.com Vol 28 May, 2023 11 Articles was performed with the mITT3 (98) population The BNT162b2 vaccine boost increased the percentage (Supplementary Table S1). of CD8+ IFN-γ+ T-cells after re-stimulation with a spike The percentage of participants with a ≥4-fold change peptide pool SA (p = 0.0089) and SB (p = 0.0018) at week in binding antibodies in the PHH-1V and BNT162b2 2 compared with baseline, showing higher values in the vaccine groups were similar, 98.4% (n = 490) and 98.8% increase rate than the PHH-1V boost (p = 0.0147 and (n = 238) respectively, on day 14 post-boost (Table 2). p = 0.0248, respectively). Moreover, the BNT162b2 Accordingly, there were non-significant differences be- group increased the percentage of CD8+ IL-4+ T-cells tween the odds of the two groups, with an odds ratio of after re-stimulation with the SB spike peptide pool at 1.31 (95% CI [0.34, 5.00]; p = 0.70). The data concerning week 2 compared with baseline (p = 0.0123). day 28 and 98 after booster administration reinforce this Concerning the exploratory endpoint of SARS-CoV-2 result, with non-significant differences in the odds be- infections ≥14 days post-booster administration, at the tween treatment groups (D28 odds ratio 1.48, 95% CI data cut-off date of the report the percentage of patients [0.47, 4.64]; p = 0.50 and D98 odds ratio 0.95, 95% CI who reported experiencing COVID-19 (n = 82) remained [0.22, 4.16]; p = 0.94). Nevertheless, the quantitative similar in the PHH-1V (10.14%) and BNT162b2 comparison of the total binding antibody titres between (11.90%) groups (OR 0.83, 95% CI [0.51, 1.36]; p = 0.45). treatments reflects that vaccination with PHH-1V None of the subjects developed severe COVID-19 and, induced higher antibody titres on days 14, 28 and 98 consequently, no hospital or intensive care unit (ICU) compared with the BNT162b2 booster dose (p < 0.0001; admission were required. Fig. 2B). These results indicate that a similar percentage of subjects respond to both vaccines with a ≥4-fold Discussion change in binding antibodies, although the subjects Our interim results from the Phase IIb HH-2 trial show vaccinated with PHH-1V reach higher titres compared that PHH-1V as a heterologous boost elicits a non- with those immunized with BNT162b2. Interestingly, inferior neutralizing antibody response to SARS-CoV-2 quantitative results on day 98 show that the decrease in Delta variant, a superior neutralizing antibody the titre in the PHH-1V arm is not as pronounced as in response against Beta variant and, most importantly, a the BNT162b2 arm, with geometric mean fold reduction superior neutralizing antibody response against the ratio from day 14 to day 98 of 1.87 (95% CI [1.35, 2.60]; Omicron variant of concern (sublineage BA.1, predom- p < 0.0001), suggesting a greater maintenance of the inant during the development of the trial) at 14 days long-term antibody titre (Fig. 2B). post-vaccination when compared to the BNT162b2 The results for the SARS-CoV-2 neutralisation test boost. Although the PHH-1V as a booster could not (VNA) for the Omicron BA.1 variant on day 14 show a demonstrate non-inferiority to BNT162b2 against the GMT for treatment ratio for BNT162b2 active control vs initial Wuhan-Hu-1 SARS-CoV-2 strain on days 14 and PHH-1V of 0.44 (95% CI [0.30, 0.65]) followed by p 28, it provided non-inferiority on day 98. Moreover, on value for ratio = 1 (p = 0.0001; Fig. 2C and Table 2). The day 98, PHH-1V vaccine showed superiority for the baseline characteristics of this subset of 58 participants Delta variant. These results suggest a lower decrease in is shown in Supplementary Table S4. the neutralizing antibody titres after the PHH-1V boost For the secondary endpoint of assessing T-cell re- when compared to BNT162b2. Accordingly, given the sponses towards the SARS-CoV-2 spike glycoprotein very high levels of community transmission observed peptides, participants in the PHH-1V group showed a for the currently circulating Delta and Omicron variants, significant increase of IFN-γ producing lymphocytes neutralisation antibody response against these variants upon in vitro re-stimulation with SARS-CoV-2 spike is of paramount importance. No significant differences peptide pools SA and RBD (4 pools covering Wuhan- were observed between the groups for those participants Hu-1, Alpha, Beta, and Delta variants; Supplementary with ≥4-fold change in binding antibodies up to 98 days material). A significant activation of CD4+ T-cells after boost administration, although the PHH-1V expressing IFN-γ was observed for the PHH-1V group at induced higher titres compared with the BNT162b2 2 weeks compared with baseline upon the stimulation boost. Our results have also shown that PHH-1V as a with the RBD (Wuhan-Hu-1, Alpha, and Beta variants) booster dose induces a significant increase of CD4+ and and Spike SA peptide pool (Fig. 4A; p < 0.0001). For CD8+ T-cells expressing IFN-γ on day 14. Concerning these variants, the increase rate in the percentage of the safety profile, PHH-1V group had less percentage of CD4+ IFN-γ+ T-cells from baseline to 2 weeks was adverse events compared with BNT162b2 group, with higher in the PHH-1V group compared to the most of mild intensity, and similar non-severe COVID- BNT162b2 group (Wuhan-Hu-1, p = 0.0097; Alpha, 19 cases. p = 0.0030; Beta, p = 0.0069). Regarding the CD8+ For the trial’s primary endpoint of humoral immu- T-cells, the boost with the PHH-1V vaccine significantly nogenicity against the ancestral Wuhan-Hu-1 strain, the increased the percentage of IFN-γ-expressing cells upon BNT162b2 booster elicited a superior neutralizing anti- re-stimulation with RBD from SARS-CoV-2 Beta variant body response on days 14 and 28, with a non-inferiority at week 2 compared to baseline (Fig. 4B; p = 0.0372). demonstrated for PHH-1V only at Day 98. This result 12 www.thelancet.com Vol 28 May, 2023 Articles a RBD Wuhan-Hu-1 RBD Wuhan-Hu-1 RBD Wuhan-Hu-1 Vaccine ns ** ns 0.12 0.3 0.09 PHH−1V 0.08 0.2 0.06 BNT162b2 0.04 0.03 0.1 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 RBD Alpha (B.1.1.7) RBD Alpha (B.1.1.7) RBD Alpha (B.1.1.7) ns ** ns 0.06 0.09 0.2 0.04 0.06 0.1 0.02 0.03 % CD4+ IFN − γ+ T cells / CD4+ T cells % CD4+ IL − 2+ T cells / CD4+ T cells % CD4+ IL − 4+ T cells / CD4+ T cells 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 RBD Beta (B.1.351) RBD Beta (B.1.351) RBD Beta (B.1.351) ns ** ns 0.12 0.15 0.3 0.09 0.10 0.2 0.06 0.05 0.1 0.03 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 Spike SA (peptide pool A) Spike SA (peptide pool A) Spike SA (peptide pool A) ns ns ns 0.06 0.15 0.3 0.04 0.10 0.2 0.02 0.05 0.1 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 Spike SB (peptide pool B) Spike SB (peptide pool B) Spike SB (peptide pool B) ns ns ns 0.06 0.15 0.20 0.04 0.10 0.15 0.10 0.02 0.05 0.05 0.00 0.00 0.00 Baseline Day 14 Baseline Day 14 Baseline Day 14 b RBD Wuhan-Hu-1 RBD Wuhan-Hu-1 RBD Wuhan-Hu-1 Vaccine ns ns ns 0.4 PHH−1V 0.02 0.15 0.3 BNT162b2 0.10 0.2 0.01 0.05 0.1 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 RBD Alpha (B.1.1.7) RBD Alpha (B.1.1.7) RBD Alpha (B.1.1.7) ns ns ns 0.15 0.6 0.03 0.10 0.4 0.02 0.01 0.05 0.2 % CD8+ IFN − γ+ T cells / CD8+ T cells % CD8+ IL − 2+ T cells / CD8+ T cells % CD8+ IL − 4+ T cells / CD8+ T cells 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 RBD Beta (B.1.351) RBD Beta (B.1.351) RBD Beta (B.1.351) ns ns ns 0.15 0.4 0.04 0.3 0.03 0.10 0.2 0.02 0.05 0.01 0.1 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 Spike SA (peptide pool A) Spike SA (peptide pool A) Spike SA (peptide pool A) ns * ns 0.8 0.09 0.6 0.4 0.06 0.4 0.2 0.03 0.2 0.00 0.0 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 Spike SB (peptide pool B) Spike SB (peptide pool B) Spike SB (peptide pool B) ns * ns 0.15 0.03 0.3 0.10 0.02 0.2 0.01 0.05 0.1 0.00 0.00 0.0 Baseline Day 14 Baseline Day 14 Baseline Day 14 Fig. 4: Characterization of T-cell responses in PBMCs from groups receiving a heterologous boosting vaccination with PHH-1V (in grey) or BNT162b2 (in blue). CD4+ T (A) and CD8+ T (B) cells were characterized by intracellular cytokine staining (ICS) of interleukin-2 (IL-2+; left panel), www.thelancet.com Vol 28 May, 2023 13 Articles may be explained by the PHH-1V vaccine being based activation of CD4+ T-cells expressing IFN-γ upon re- on a fusion heterodimer consisting of the spike RBD stimulation with pools of RBD peptides from different sequence from the SARS-CoV-2 Beta and Alpha vari- variants. Interestingly, this response is stronger ants, while the BNT162b2 vaccine encodes the SARS- compared to those subjects boosted with BNT162b2 CoV-2 full-length spike protein of Wuhan-Hu-1.30 vaccine, indicating that the PHH-1V heterologous boost For the primary endpoint of safety and tolerability, is more efficient than the homologous boost in the the percentage of subjects with local and systemic re- activation of the CD4+ T-cell memory previously actions, total adverse effects and treatment-related induced by the primary vaccination protocol. As no IL-4 adverse events were lower in the PHH-1V group expression was detected in the activated CD4+ T-cells compared to the BNT162b2 group. No deaths were re- after the in vitro re-stimulation, the ICS results suggest ported in the study in either of the PHH-1V or that the PHH-1V booster induces a Th1-biased T-cell BNT162b2 group and one SAE was reported, which was response. The detection of the IFN-γ expression by the resolved and found to be nonrelated to the vaccination. ELISpot assay confirms the Th1-biassed T-cell response As of day 98, a total of 82 SARS-CoV-2 cases were re- upon boost immunization with the PHH-1V vaccine. ported, and no statistically significant differences were Avoiding a Th2-biased immune response is important observed between the PHH-1V booster and the active as it has been related with ineffective vaccines that control, BNT162b2. In addition, no severe cases of induce vaccine-associated enhanced respiratory disease COVID-19 were reported in any of the vaccinated (VAERD) after subsequent infection.33 Moreover, the groups. Of note, this study took place temporarily heterologous boost with the PHH-1V vaccine has during a period of high incidence of COVID-19, namely proven to induce the activation of CD8+ T-cells the currently circulating Delta and Omicron (BA.1 expressing IFN-γ. The T-cell response is crucial to variant) waves in Spain. Moreover, most of participants confer protection from COVID-19 severe disease,4 and in the study were healthcare workers, thus with a these results show the potential ability of PHH-1V not higher infection exposition compared to the general only to fight against the virus with neutralizing anti- population.31 bodies but also with cellular immunity specifically to When assessing the secondary endpoint of humoral destroy infected cells. immunogenicity against different variants of SARS- When comparing PHH-1V with other vaccines, the CoV-2, remarkably, PHH-1V was able to sustain a CoVLP is a plant-produced virus-like particle (VLP) good neutralising ability against the Omicron BA.1 vaccine under development, and the CoV2 preS dTM is SARS-CoV-2 variant despite the heavily mutated spike an authorised recombinant protein antigen vaccine protein of this variant, many of them located in the based on an S-protein sequence from the Wuhan-Hu-1 RBD.32 This fact strongly supports the high humoral (D614) reference strain.34,35 Both candidate vaccines immunogenicity of the PHH-1V RBD-based candidate have been initially tested without adjuvant, but elicited against a wide range of potential new mutations since better immunogenicity results with the AS03 adjuvant the PHH-1V antigen comprises key mutations that are (GlaxoSmithKline Vaccines, Rixensart, Belgium), and also present in the Omicron BA.1 variant, as well as in are now called CoVLP-AS03 and CoV2 preS dTM-AS03, many other SARS-CoV-2 variants. In addition, our re- respectively. In addition, their phase II results indicate sults in humoral immunogenicity were confirmed when that they are well tolerated and have favourable all the subjects who reported COVID-19 infections were neutralizing antibody titres, with both being stable at excluded of the data analysis, thus eliminating the refrigerator temperatures. Unlike PHH-1V, which is possible bias of an uneven distribution of infection- being tested as a booster only, these vaccines were tested induced immunogenicity. for primary (2-dose) vaccination schedules, thereby, For the secondary endpoint of T-cell-mediated without non-inferiority assessments with already response determined by ELISpot, the results indicate approved vaccines. The NVX-CoV2373 subunit vaccine that the PHH-1V boost, after a primary vaccination with was also tested for a 2-dose vaccination schedule, elic- BNT162b2, increases the cellular immune response af- iting lower IgG levels compared to the BNT162b2 and ter the in vitro re-stimulation. The ICS data for T-cell mRNA-1273 mRNA vaccines and a poor induction of response characterization demonstrates that the booster spike-specific CD8+ T-cells.36,37 These data contrast with immunization with the PHH-1V vaccine induces the those reported in the present study for PHH-1V. interferon gamma (IFN-γ+; middle panel) and interleukin-4 (IL-4+, right panel). Boxes depict the median (solid line) and the interquartile range (IQR), and whiskers expand each box edge 1.5 times the IQR. Interaction contrasts have been displayed in the plots, comparing the increase rates over time between the two vaccination groups. Statistically significant differences are shown as * for p ≤ 0.05; ** for p ≤ 0.01; *** for p ≤ 0.001. Non-significant comparisons have been indicated with “ns”. p-values lower than 0.05 indicate that one of the vaccinated groups has experienced a stronger boost compared to the other. RDB; receptor binding domain for the SARS-CoV-2 spike protein (ancestor Wuhan-Hu-1 strain); RDB B.1.1.7 (Alpha variant); RDB B.1.351 (Beta variant); Spike SA corresponds to 194 spike protein peptide pools overlapping the S1-2016 to S1-2196 region of the Spike protein; Spike SB corresponds to 168 spike protein peptide pools overlapping the S1-2197 to S2-2377 region of the Spike protein. 14 www.thelancet.com Vol 28 May, 2023 Articles However, the NVX-CoV2373 induced spike-specific Declaration of interests CD4+ T-cells to all tested VOCs (Wuhan-Hu-1, Alpha, The authors of this manuscript declare: J Blanco has received institu- tional grants from HIPRA, Grifols, and MSD, royalties for licensed Beta, Delta and Omicron BA.1 and BA.2).37 The CoVLP- patent from AlbaJuna, honoraria for lectures from FLS Science and AS03, CoV2 preS dTM-AS03, and NVX-CoV2373 vac- CIBER, supporting for meeting and/or travel from Gilead Sciences, and cines are now also being considered as booster doses for unpaid independent COVID-19 monitoring from GMCSC (Multidisci- primary vaccination schedules. plinary Collaborative Group for the Scientific Monitoring of COVID-19) The main limitations of this study are related to the and unpaid participation in COVID-19 advisory group for CCAC (Comitè Científic Assessor de la COVID-19). Outside of this work, J changing epidemiological and social situation during the Blanco is the CEO, founder and shareholder of AlbaJuna Therapeutics, pandemics. Although it was initially proposed that neu- S.L. J Corominas, C Garriga, A Barreiro, L González-González, L tralising antibodies by means of PBNA would have to be Madrenas, I Güell, D Raïch-Regué, J G Prado, T Prat, E Torroella, B determined for Wuhan-Hu-1 strain and Alpha, Beta and Trinité, L Ferrer, M Cañete and A Prenafeta have received funding from HIPRA. The funding from HIPRA to R Ramos was paid to his insti- Delta SARS-CoV-2 variants, the situation of the pan- tution. A Soriano has received grants from Pfizer and Gilead Sciences, demics during the study made us change the proposal consulting fees from Pfizer, MSD and Shionogi, and honoraria for and so Alpha was substituted for Omicron which was the lectures for Pfizer, MSD, Gilead Sciences, Shionogi, Angelini, and relevant circulating variant at that time, adding a very Menarini. B Trinité declares royalties by an institutional agreement and important value to the study that has demonstrated the consulting fees for HIPRA, and is an unpaid member in advisory board for the Health Department of the Generalitat de Catalunya. N Izquierdo- effective potential of the PHH-1V vaccine as a heterolo- Useros, D Raïch-Regué and M Gallemí declare institutional grants from gous booster. However, pseudovirions against the more HIPRA, Pharma Mar, Grifols, Dentaid, Palobiofarma, Mynorix and recent circulating SARS-CoV-2 variants, such as Omi- Amassence. N Izquierdo-Useros and M Gallemí have received speaking cron BQ.1.1 or XBB, for example, are not yet available to honoraria from FLS Science. JR Arribas has received consulting fees and payment for participating in advisory board from Gilead Sciences, MSD, perform most of the assays presented in this article. GSK, Eli Lilly, Roche, Pfizer and Sobi, honoraria for lectures and sup- In conclusion, the HH-2 study demonstrates that the port for meetings and/or travel from MSD. A Borobia has received PHH-1V vaccine is safe and well tolerated, eliciting a grants from GSK, Moderna and Janssen, speaking honoraria for Jans- strong neutralizing antibody response against all tested sen, Gilead Sciences and Pfizer, and payment for participating in advi- SARS-CoV-2 variants and Wuhan-Hu-1 strain. Although sory board for Pfizer, Janssen and MDI. PM García has received consulting fees and speaking honoraria from Gilead Sciences, Mundi- PHH-1V did not reach non-inferiority at day 14 and day pharma and Pfizer, payment for expert testimony and participated in 28 in neutralizing antibody response against the advisory board for Gilead Sciences and received support for meeting Wuhan-Hu-1 strain, it did so at day 98. The study spe- and/or travel from Pfizer. S Otero-Romero has received speaking hon- cifically shows superiority against the Beta and Omicron oraria from Genzyme, Biogen-Idec, Novartis, Roche, and MSD. Julia G Prado declares institutional grants from Grifols. BA.1 SARS-CoV-2 variants when compared to J Corominas, C Garriga, A Prenafeta, A Moros, M Cañete, A Bar- BNT162b2 homologous vaccine boost and non- reiro, L González-González, L Madrenas, I Güell, T Prat, E Torroella and inferiority against Delta on day 14. Importantly, the L Ferrer are employees of HIPRA. Some of these authors may have PHH-1V vaccine also allows for better sustained levels stocks of HIPRA. of antibodies over time (day 98). This demonstrates that Several patent applications have been filed by HIPRA SCIENTIFIC S.L.U. and Laboratorios HIPRA, S.A. on different SARS-CoV-2 vaccine PHH-1V can elicit protection against current circulating candidates and SARS-CoV-2 subunit vaccines, including the novel re- variants of concern and, most importantly, can antici- combinant RBD fusion heterodimer PHH-1V. A Barreiro, J Corominas, pate protection against potentially new emerging vari- A Prenafeta, L González-González, L Madrenas, L Ferrer, E Torroella, T ants. The PHH-1V vaccine also induces a strong and Prat and C Garriga are the inventors of these patent applications. N Izquierdo-Useros is a patent inventor with no economical compensation balanced T-cell response against SARS-CoV-2. At pre- for Pharma Mar and Mynorix. sent, a phase III trial is ongoing where the inclusion The other authors have no relevant conflicts of interest to declare. criteria have been opened further and studies on the vaccine are continued. The interim results presented Acknowledgements here provide data on the balance observed between Medical writing support was provided by Vanessa Chiganças at Dy- namic Science S.L.U. (Evidenze Clinical Research, Madrid, Spain) safety and immunogenicity elicited by the PHH-1V during the preparation of this paper, and funded by HIPRA SCIENTI- boost. These data support PHH-1V vaccine as a valu- FIC, S.L.U. able tool to the current authorized COVID-19 vaccines We especially acknowledge the following members of Veristat, who and as a booster dose in the vaccination campaigns. contributed to the success of this trial. The following were responsible for study management, biostatistics, medical monitoring, data man- Contributors agement, and database programming of the study: Robin Bliss, PhD, Veristat was responsible for managing the data. Authors contributed to Emma Albacar, MPH, Nancy Hsieh, MPH, Montse Barceló, MD, the acquisition, analysis, and/or interpretation of data. All authors had Mariska van der Heijden, MSc, Amy Booth, Edmund Chiu, Avani Patel, full access to all the data, revised the manuscript critically for important and Cesar Wong, MD. intellectual content, approved the version to be published, and accepted We are grateful to Daniel Perez-Zsolt and Jordana Muñoz-Basagoiti responsibility for publication. for their outstanding contribution to VNA. The authors would like to thank Silvia Marfil and Raquel Ortiz for Data sharing statement technical assistance. All data relevant to the study are included in the article or uploaded as The authors thank Ruth Peña for technical assistance with sample supplementary information. Further data are available from the authors management and ELISpot and Gabriel Felipe Rodriguez-Lozano for upon reasonable request and with permission of HIPRA S.A. ELISpot database generation. www.thelancet.com Vol 28 May, 2023 15 Articles We would like to express our gratitude to Marina Machado, Ana 12 Mathieu E, Ritchie H, Ortiz-Ospina E, et al. A global database of Álvarez-Uría, Sara Rodríguez, Ma Jesús Pérez Granda, Juan Carlos COVID-19 vaccinations. Nat Hum Behav. 2021;5(7):947–953. López Bernaldo de Quirós, Ma Teresa Aldamiz, Francisco Tejerina, 13 Puhach O, Adea K, Hulo N, et al. Infectious viral load in unvacci- nated and vaccinated individuals infected with ancestral, Delta or Cristina Díez, Iván Adán, Ana Mur, Félix García and Víctor Fernández Omicron SARS-CoV-2. Nat Med. 2022;28:1491. for their tireless effort and contribution to this important public health 14 Agrati C, Capone S, Castilletti C, et al. Strong immunogenicity of clinical trial. heterologous prime-boost immunizations with the experimental We would like to thank Glòria Pujol and Eduard Fossas for their vaccine GRAd-COV2 and BNT162b2 or ChAdOx1-nCOV19. NPJ assistance in the revision of the manuscript; Fiorella Gallo, Núria Vaccines. 2021;6(1):131. Fuentes and Miriam Oria for the ELISA analysis; Clara Panosa, Thais 15 Andersson NW, Thiesson EM, Laursen MV, Mogensen SH, Kjær J, Pentinat and Ester Puigvert for their assistance in the production of the Hviid A. Safety of heterologous primary and booster schedules with vaccine and Jordi Palmada and Eva Pol for carrying out manufacturing ChAdOx1-S and BNT162b2 or mRNA-1273 vaccines: nationwide cohort study. BMJ. 2022;378:e070483. controls. And of course, we would like to especially thank all the HIPRA 16 Atmar RL, Lyke KE, Deming ME, et al. Homologous and heterol- workers who in one way or another have contributed to making this ogous Covid-19 booster vaccinations. N Engl J Med. 2022;386(11): project a reality. 1046–1057. The authors thank the members of the DSMB for their expertise 17 Interim recommendations for heterologous COVID-19 vaccine sch- and recommendations. edules: interim guidance, 16 December 2021. CC BY-C-A 3.0 IGO. We are indebted to the HCB-IDIBAPS Biobank, integrated in the Geneva: World Health Organization; 2021. Spanish National Biobanks Network, for the biological human samples 18 Costa Clemens SA, Weckx L, Clemens R, et al. Heterologous versus and data procurement. homologous COVID-19 booster vaccination in previous recipients of two doses of CoronaVac COVID-19 vaccine in Brazil (RHH-001): Finally, we want to express our gratitude to all the volunteers for a phase 4, non-inferiority, single blind, randomised study. Lancet. their time and effort. With their contribution they have enabled the 2022;399(10324):521–529. generation of medical and scientific knowledge that will enable us to 19 Garg I, Sheikh AB, Pal S, Shekhar R. Mix-and-Match COVID-19 draw closer to the end of this pandemic. vaccinations (heterologous boost): a review. Infect Dis Rep. 2022; Funding: This work was supported by HIPRA SCIENTIFIC, S.L.U 14(4):537–546. (HIPRA) and partially funded by the Centre for the Development of 20 Jara A, Undurraga EA, Zubizarreta JR, et al. Effectiveness of ho- Industrial Technology (CDTI, IDI-20211192), a public organisation mologous and heterologous booster doses for an inactivated SARS- answering to the Spanish Ministry of Science and Innovation. CoV-2 vaccine: a large-scale prospective cohort study. Lancet Glob Health. 2022;10(6):e798–e806. 21 Kleanthous H, Silverman JM, Makar KW, Yoon IK, Jackson N, Appendix A. Supplementary data Vaughn DW. Scientific rationale for developing potent RBD-based Supplementary data related to this article can be found at https://doi. vaccines targeting COVID-19. NPJ Vaccines. 2021;6(1):128. org/10.1016/j.lanepe.2023.100613. 22 Kyriakidis NC, López-Cortés A, González EV, Grimaldos AB, Prado EO. SARS-CoV-2 vaccines strategies: a comprehensive re- view of phase 3 candidates. NPJ Vaccines. 2021;6(1):28. 23 Barreiro A, Prenafeta A, Bech-Sabat G, et al. Preclinical evaluation References of a COVID-19 vaccine candidate based on a recombinant RBD 1 Salamanna F, Maglio M, Landini MP, Fini M. Body localization of fusion heterodimer of SARS-CoV-2. iScience. 2023;26:106126. ACE-2: on the trail of the keyhole of SARS-CoV-2. Front Med. 24 Moros A, Prenafeta A, Barreiro A, et al. Immunogenicity and safety 2020;7:594495. in pigs of PHH-1V, a SARS-CoV-2 RBD fusion heterodimer vaccine 2 Yang Y, Du L. SARS-CoV-2 spike protein: a key target for eliciting candidate. bioRxiv. 2023:2023.01.19.524684. persistent neutralizing antibodies. Signal Transduct Target Ther. 25 Prenafeta A, Bech-Sàbat G, Moros A, et al. Preclinical efficacy, safety, 2021;6(1):95. and immunogenicity of PHH-1V, a second-generation COVID-19 3 Khoury DS, Cromer D, Reynaldi A, et al. Neutralizing antibody vaccine, in non-human primates. bioRxiv. 2022:2022.12.13.520255. levels are highly predictive of immune protection from 26 Wang Y, Liu C, Zhang C, et al. Structural basis for SARS-CoV-2 symptomatic SARS-CoV-2 infection. Nat Med. 2021;27(7): Delta variant recognition of ACE2 receptor and broadly neutral- 1205–1211. izing antibodies. Nat Commun. 2022;13(1):871. 4 GeurtsvanKessel CH, Geers D, Schmitz KS, et al. Divergent SARS- 27 Chen J, Wang R, Gilby NB. Omicron variant (B.1.1.529): infectivity, CoV-2 Omicron-reactive T and B cell responses in COVID-19 vac- vaccine breakthrough, and antibody resistance. J Chem Inf Model. cine recipients. Sci Immunol. 2022;7(69):eabo2202. 2022;62(2):412–422. 5 Ramasamy MN, Minassian AM, Ewer KJ, et al. Safety and immu- 28 Leal L, Pich J, Ferrer L, et al. Safety and Immunogenicity of a Re- nogenicity of ChAdOx1 nCoV-19 vaccine administered in a prime- combinant Protein RBD Fusion Heterodimer Vaccine against boost regimen in young and old adults (COV002): a single-blind, SARS-CoV-2: preliminary results of a phase 1-2a dose-escalating, randomised, controlled, phase 2/3 trial. Lancet. 2021;396(10267): randomized, double-blind clinical trial. medRxiv. 2022, 2022.08. 1979–1993. 09.22278560. 6 Stuart ASV, Shaw RH, Liu X, et al. Immunogenicity, safety, and 29 Pradenas E, Trinité B, Urrea V, et al. Stable neutralizing antibody reactogenicity of heterologous COVID-19 primary vaccination levels 6 months after mild and severe COVID-19 episodes. Med (N incorporating mRNA, viral-vector, and protein-adjuvant vaccines in Y). 2021;2(3):313–320.e4. the UK (Com-COV2): a single-blind, randomised, phase 2, non- 30 Walsh EE, Frenck RW Jr, Falsey AR, et al. Safety and immunoge- inferiority trial. Lancet. 2022;399(10319):36–49. nicity of two RNA-based Covid-19 vaccine candidates. N Engl J Med. 7 Zhu FC, Li YH, Guan XH, et al. Safety, tolerability, and immu- 2020;383(25):2439–2450. nogenicity of a recombinant adenovirus type-5 vectored COVID-19 31 Lucaccioni H, Costa C, Duque MP, Balasegaram S, Sá Machado R. vaccine: a dose-escalation, open-label, non-randomised, first-in- Risk of COVID-19 in health professionals: a case-control study, human trial. Lancet. 2020;395(10240):1845–1854. Portugal. Port J Public Health. 2021;39(3):137–144. 8 Heinz FX, Stiasny K. Distinguishing features of current COVID-19 32 Cheng SMS, Mok CKP, Leung YWY, et al. Neutralizing antibodies vaccines: knowns and unknowns of antigen presentation and against the SARS-CoV-2 Omicron variant BA.1 following homolo- modes of action. NPJ Vaccines. 2021;6(1):104. gous and heterologous CoronaVac or BNT162b2 vaccination. Nat 9 Fahrni ML, Ismail IA, Refi DM, et al. Management of COVID-19 Med. 2022;28(3):486–489. vaccines cold chain logistics: a scoping review. J Pharm Policy 33 Graham BS. Rapid COVID-19 vaccine development. Science. Pract. 2022;15(1):16. 2020;368(6494):945–946. 10 Dolgin E. How protein-based COVID vaccines could change the 34 Hager KJ, Pérez Marc G, Gobeil P, et al. Efficacy and safety of a pandemic. Nature. 2021;599(7885):359–360. recombinant plant-based adjuvanted Covid-19 vaccine. N Engl J 11 McDonald I, Murray SM, Reynolds CJ, Altmann DM, Boyton RJ. Med. 2022;386(22):2084–2096. Comparative systematic review and meta-analysis of reactogenicity, 35 Sridhar S, Joaquin A, Bonaparte MI, et al. Safety and immunoge- immunogenicity and efficacy of vaccines against SARS-CoV-2. NPJ nicity of an AS03-adjuvanted SARS-CoV-2 recombinant protein Vaccines. 2021;6(1):74. vaccine (CoV2 preS dTM) in healthy adults: interim findings from a 16 www.thelancet.com Vol 28 May, 2023 Articles phase 2, randomised, dose-finding, multicentre study. Lancet Infect 37 Hielscher F, Schmidt T, Klemis V, et al. NVX-CoV2373-induced Dis. 2022;22(5):636–648. cellular and humoral immunity towards parental SARS-CoV-2 36 Heath PT, Galiza EP, Baxter DN, et al. Safety and efficacy of NVX- and VOCs compared to BNT162b2 and mRNA-1273-regimens. CoV2373 Covid-19 vaccine. N Engl J Med. 2021;385(13):1172–1183. J Clin Virol. 2022;157:105321. www.thelancet.com Vol 28 May, 2023 17