This is an important study that compares heterologous (different vaccines for the 1st and 2nd doses) with homologous schedules (same vaccine for the 1st and 2nd dose) with regard to immunogenicity, safety and reactogenicity.   It takes two groups of individuals who have previously received either the AZ or BioNTech vaccines and randomises those individuals in each group to one of three 2nd doses – Moderna, Novavax or the corresponding homologous dose (that is, either AZ or BioNTech).  Since the 1st dose was not assigned randomly, there can be no causal comparison between AZ and BioNTech, and so the comparisons correctly reported in the paper are within each group.  That is, in each case the comparisons are between the Moderna or Novavax 2nd dose with the reference homologous schedule.  

The primary endpoint is the geometric mean ratio (GMR) of IgG (antibody) concentration 28 days after the 2nd dose, where a ratio of 1 represents the case were the heterologous and homologous schedules generate the same antibody levels. The study is designed to show non-inferiority of the heterologous schedule to the homologous schedule.  That is, that mixed dosing is no worse than receiving the same vaccine both times. It study uses a criteria of 0.63 as the non-inferiority margin, which means that so long as the lower 98.75% confidence limit (one-sided) of the GMR is above 0.63, then non-inferiority of the heterologous schedule can be claimed.    

The top level result is that both AZ/Moderna and AZ/ Novavax were no worse than AZ/AZ in terms of IgG concentration 28 days post 2nd dose, whereas only BioNTech/Moderna was no worse than BioNTech/BioNTech.  However the BioNTech/Novavax combination did not meet the criteria since the lower confidence limit of 0.4 was <0.63.

It is also possible to look at superiority within each group once non-inferiority has been evaluated, and the authors report these results also. In this case, one looks at the same lower confidence limit for the GMR and compares it to 1 instead of 0.63.  Looking at it in this way showed that the heterologous dosing of both AZ/Moderna and AZ/ Novavax were superior to AZ/AZ while BioNTech/Moderna was also shown to be superior to BioNTech/BioNTech.  In particular the AZ/Moderna combination showed a sizeable increase in IgG versus AZ/AZ with a lower confidence limit for the GMR of 8.4.  This is consistent with the earlier results from the University of Oxford-led Com-Cov study (1) of the four AZ and BioNTech (i.e. mRNA) two-dose combinations, that provided strong evidence that AZ/BioNTech 4 week dosing provided a better immune response than AZ/AZ (lower 97.5% confidence limit of 7.5 for the corresponding GMR).

It is tempting to try to compare the IgG concentrations between AZ and BioNTech, but since individuals were not randomised to the 1st dose, this could be highly misleading, despite the demographics being very similar in the two groups, since there may be confounders. Indeed, the authors note a higher prevalence of co-morbidity (cardiovascular, respiratory and diabetic) in the BioNTech group, reflecting the earlier availability of the BioNTech vaccine in the UK.

Overall the study supports the view that mixing vaccines is a sensible approach in terms of generating a higher antibody response than would otherwise be achieved.  The only combination shown to be not non-inferior was the BioNTech/Novavax combination, although the Novavax vaccine is not yet approved in the UK, and as such it not part of the vaccine roll out.  Although 15 serious adverse events were reported, none were classified as being related to the vaccines, which provides further evidence for the safety of all the vaccines studied when used alone or in combination.

1. Liu X, Shaw RH, Stuart ASV et al.  Safety and Immunogenicity Report from the Com-COV Study – a Single-Blind Randomised Non-Inferiority Trial Comparing Heterologous And Homologous Prime-Boost Schedules with An Adenoviral Vectored and mRNA COVID-19 Vaccine (Com-COV).  Lancet 2021; 398;856-69.

‍

This paper reports the results of a study investigating a ‘mix and match’ approach to a two dose primary vaccination course and demonstrate, as expected from first principles, the superiority of mixing two different vaccines (heterologous vaccination) to repeat dosing with the same vaccine (homologous vaccination) in terms of the antibody and cellular response generated following the second dose. The study also demonstrates the very high level of antibody response following an mRNA vaccine when given as the second dose in a heterologous mixture, when compared to either a viral vector (ChAdOx01) or a protein (Nvx) vaccine. In the absence of agreed correlates of protection against infection, illness and death with these vaccines, there is a tendency to use vaccine schedules which offer potential for a greater immune response. Based on this presumption, the immune response data suggests heterologous vaccination might have some benefits to offer in this respect, particularly for subjects receiving the viral vector vaccination as the first dose. Not surprisingly, the improved immunological response is accompanied by a greater potential for reactogenic effects, particularly systemic effects, occurring most frequently following a second dose with a mRNA vaccine. One missing piece of information in this study was the impact of heterologous vaccination using the protein subunit vaccine for the first of two doses and it would be of interest to have that information. In practical terms, these data demonstrate that, in a world where supply continues to fall behind demand, health care personnel can use any combination of vaccines available within their country to ensure completion of a primary course. In addition, individuals that have an allergic reaction to the first vaccine dose can be given an alternative vaccine type without, for the most part, losing the benefit of a second dose. The same may be true for continued boosting with heterologous vaccine types, where necessary although this requires further investigation as more vaccines approach the market

Speaking on behalf of CEPI, a co-funder of the Com-COV2 trial:

We’re pleased to have co-funded this crucial area of research, in collaboration with our partners at the University of Oxford and the UK Government, supporting excellent science conducted in the UK that advances vaccine research for the benefit of all. This is yet another example of the impact that funding innovative R&D can have on our hopes of ending the COVID-19 pandemic.

With COVID-19 cases continuing to rise and the emergence of new variants like Omicron, it is imperative that we rapidly protect as many people as possible from this devastating virus. As has long been said, no one is safe until everybody is safe, and we hope that today’s trial findings will contribute to our work to achieve this vital goal.

This is extremely encouraging and valuable data on the potential to mix-and-match COVID-19 vaccines in primary immunisation schedules. Knowing that a second dose of a different COVID-19 vaccine can generate a robust immune response is advantageous in helping the rollout of COVID-19 vaccines through COVAX, especially in populations still urgently waiting for their primary immunisation or in those partially vaccinated.

‍