FormalPara Key Summary Points

Influenza and associated complications can cause a significant burden of disease beyond respiratory illness.

There is a need for randomized studies in both clinical and real-world settings as well as consistency across results to obtain high-quality evidence.

Clinical outcome assessment should include important outcomes such as medically attended disease and hospitalization rates.

The high-dose influenza vaccine has been shown to have considerable effectiveness in older adults, especially in vulnerable populations.

Podcast: Need for quality evidence for decision-making on seasonal influenza vaccine (MP4 514220 KB)

Digital Features

This article is published with digital features, including podcast audio to facilitate understanding of the article. To view digital features for this article, go to https://doi.org/https://doi.org/10.6084/m9.figshare.25075439.

Podcast Transcript

Podcast Attendees and Introduction

Ann: Welcome to the podcast on Need for Quality Evidence for Decision-Making on Seasonal Influenza Vaccines. My name is Dr. Ann Falsey, I’m a Professor of Medicine at the University of Rochester. And I’m joined by two experts.

Stefania: Hi. I’m Stefania Maggi, geriatrician from the Research Council of Italy based in Padua, and I’m Research Director of the Aging Branch.

Ann: And Dr. Biering-Sorensen.

Tor: Yeah, thank you so much, Dr. Falsey. My name is Tor Biering-Sorensen and I’m a cardiologist from Copenhagen, Denmark. I’m a professor in pragmatic trials at University of Copenhagen.

Ann: Well, thank you.

Impact of Influenza Beyond the Respiratory System

Ann: So we’ll get started. And one of the things that we want to discuss today is the need for quality evidence. And there’s a lot to be discussed on this topic because not only are we interested in the traditional outcomes of influenza such as acute respiratory illness, there’s now quite a bit of evidence of indirect effects of influenza [1,2,3]. Some of the things we’re considering when we talk about older adults are the functional complications that occur with flu [2, 3]. In addition, there are some very intriguing data on the neurological effects of flu, both the links with Alzheimer’s and the possibility that influenza vaccine may be protective, and clearly we need good data there [1]. So, I’ll turn to you, Dr. Maggi, and just ask the question, what are some of the other major organ systems that can be impacted by influenza infection? Not just limited to the respiratory tract.

Stefania: Right. And I think that we should underline the fact that influenza is more of a systemic disease, not just limited to the respiratory system as we have complications that can range from mild to severe and might affect different organs [4,5,6,7,8]. Can be short-term or long-term. And of course, the best-known short-term complication is pneumonia where we know that there is up to a tenfold increase in the risk of developing pneumonia after an influenza episode [6,7,8,9]. But we also know that there are serious cardiovascular complications, such as heart attack, heart failure [4, 7, 8, 10, 11]. We also have people with chronic diseases such as people with diabetes that have an increased risk of developing hyperglycemia [5]. And we know that hyperglycemia is itself a risk factor for infectious diseases. So, they start a vicious cycle that might be very negative with complications involving also micro and macrovascular complications [5]. So, with regard in general with the medium, long-term risk, we know that comorbidity is very frequent in older adults. And the exacerbation of these chronic conditions such as COPD, heart disease, diabetes as we mentioned, it’s very frequent, and therefore, very negative outcome well beyond the respiratory system only [4,5,6, 8, 9, 11].

Ann: Well, thank you. And certainly, the issue of strokes and heart attacks feeds into the narrative that there can be loss of function [4,5,6,7,8].

Challenges in Quantifying the Benefits of Influenza Vaccination

Ann: So, let me continue by asking, what are the challenges in quantifying the benefits of influenza vaccination? And maybe I’ll start, again, with you, Dr. Maggi, to give an overview of what the elements are that should be assessed when quantifying vaccine benefits.

Stefania: Yes. So, we have several elements, and of course, immunogenicity that is required for registration purposes is not enough because we want to have data on the clinical outcomes. So, the reduction in influenza-related morbidity and mortality, therefore, the number and the severity of influenza-related illness, hospitalizations, deaths are all crucial elements. And the health care utilization in general also with the emergency room visits, the outpatient clinic visits for influenza-related illness is very important. And on top of that, we must always make sure that we include the selected at-high-risk populations such as older individuals, pregnant women, children, and all those with underlying health conditions that are considered high-risk [12, 13]. The other elements from the methodological point of view is that we must be sure that we have data that are consistent across different studies, different populations, and different seasons [13]. And also, the design, of course, would require randomized studies, both in the clinical and in the real-world settings compared to the standard of care [12, 13].

Ann: Well, great. So, it sounds easy, right? So, Dr. Biering-Sorensen, what would you describe as some of the challenges associated with achieving those outcomes?

Tor: Some problems exist if we only rely on immunogenicity trials alone because the evidence from those kind of trials are simply not sufficient because the fact that a vaccine might lead to an immune response, it does not translate into whether or not a vaccine protects the individual that is exposed to a pathogen. And of course, the immunogenicity also varies depends on age and also subsequent exposure to antigens and comorbidities and so on across several seasons, and also using different study designs and the thing that we at least in my country have done with our trials is that we focus on hard outcomes as our primary outcomes [12]. So that’s hospitalization with lab confirmed cases of flu or pneumonia making the outcomes that we assess more relevant also when conducting cost-effectiveness analysis, and afterwards, it’s actually whether we prevent hospitalizations, which we know is some of the most important parts of a patient’s trajectory in the healthcare system [6]. So, using additional outcomes than just only looking at how the immune system responds to vaccines. And again, we need to underscore that if we only rely on observational data, meaning that relying on data that is not obtained from randomized trials, we know that the vaccine effectiveness that we assess in these kind of studies, they are prone to biases that may not give us a true vaccine effectiveness [12, 14]. Hence, we need randomized trials to assess what the vaccine effectiveness is, and that is also important to underscore in this discussion that we’re having today. In the conventional, original randomized trials, we are always limited by the very long list of inclusion and exclusion criteria making our patient population as part of the patient characteristics as published in the trial manuscript [15]. So, conducting controlled trials under very controlled circumstances with very long lists of inclusion and exclusion criteria makes it less generalizable. When conducting trials in this way, we also reduce the barriers of some of the trial participants that could have participated in trials that normally are not seen as trial participants [12].

Quality of Evidence and Study Design

Ann: Well, thank you. And you’ve touched on some of the things that we’d like to discuss next. And so, I’ll address this question to both of you, and that is, how does the study design impact the quality of the evidence in clinical trials that assess both efficacy and safety of influenza vaccine? So, I’ll let Dr. Maggi begin, and then you can finish up maybe telling us about some of, again, your creative pragmatic trials that you’ve conducted. So, Stefania?

Stefania: What we do want to know is that the efficacy measures how well a vaccine prevents a specific outcome in a selected study population, and these are studies that are conducted usually under very tightly controlled conditions, often with strict inclusion and exclusion criteria [15]. And as geriatrician, we know very well that very often, the most vulnerable, frail, comorbid older adults are excluded by these clinical trials [14]. Therefore, the effectiveness to the contrary is the evaluation of how vaccines have the ability to prevent outcomes in routine everyday conditions, conducted, therefore, in real-world settings [14,15,16]. And often, they use observational study designs that unfortunately are subject to all the biases that have been mentioned before and the confounding factors that only, with randomization, we can control. Why? Because the randomization of the patients guarantees that we have the same possibility of allocating the patients in one of the two groups: One of the vaccines under evaluation, and, let’s say, the standard dose vaccine [14,15,16] So the randomization can guarantee that we avoid the selection bias, and we have high likelihood that the characteristics influencing the outcomes are evenly distributed between the two groups, therefore, they are comparable at baseline [14,15,16]. And also, the random assignment helps to control for the known and unknown confounding that might affect the study outcomes. Therefore, we distribute these variables randomly across groups and we can be sure that the differences that we will see are not due to any imbalance baseline among the two groups [14, 16]. Finally, this randomization allows us to be sure that we have a representative sample of the real population, and therefore, we can generalize our results [14,15,16].

Ann: Yeah. So, Tor, can you mention a little bit about some of your pragmatic and your creative solutions to some of these issues?

Tor: In Denmark, like in other Scandinavian countries, we have nationwide registries where all contact with our healthcare system is registered because we have this free-of-charge healthcare system which is paid for through our taxes. But another unique thing that we have in Denmark is that we have an electronic emailing system that is only used for priority mail. And this email system is linked to these nationwide registries through our Social Security Number. So, we can identify a potential trial population through the nationwide registries. For example, if for our flu trials that we are doing currently, we invite all Danish citizens 65 years and above to participate in our trial called DANFLU-2 where we are randomizing close to 300,000 trial participants to either receiving a high-dose or standard dose flu vaccine [17, 18]. And the unique thing is that besides leveraging the nationwide registries to identify all the 1 million participants that we know can participate in our trial in Denmark, we also leverage the system and the e-Boks system to invite all of these 1 million Danish citizens to participate in our DANFLU-2 trial [18]. And this system is the reason why it’s possible for us to randomize close to 300,000 citizens within a nationwide pragmatic trial testing two different flu vaccines. And this is only possible because we, of course, leverage the nationwide registries, leverage the e-Boks system, but also because we have a healthcare system that is completely digitalized, making it possible for us to invite so many citizens through this e-Boks system, but also making it possible for us to obtain all baseline characteristics through our nationwide registries, and making it possible for us to obtain all outcomes. So, all hospitalizations and mortality through the follow-up period through the nationwide registries. So, I think the digitalization of healthcare systems throughout the globe will be the future for the possibility of conducting these giga-trials which are sufficiently powered to assess whether vaccines reduce hard outcomes and whether it makes sense for a health authority to implement vaccines on a population level to potentially reduce the cost of hospitalizations.

Ann: Well, thank you. And have to say that many of us in other parts of the world are very envious of your system because as you mentioned, randomization is really critical to obtaining good quality evidence, and certainly we’ll have to learn a lot from the Danes because you can do these very large studies and you can do them in a randomized fashion, which gives us the best data.

Innovative Patient Randomization

Ann: Tor, are there any other creative or innovative trials that you’re aware of that are ongoing that address these questions?

Tor: Yeah. I actually know of a study conducted within the Kaiser system where they conducted a very large-scale cluster randomized trial of 1.6 million people in the US randomizing on which vaccine that administered at which day. And they compared the high-dose recombinant flu vaccine versus the standard dose on lab-confirmed influenza cases. And as I mentioned, Ann, it was randomized—1.6 million. So that’s a huge number of trial participants [19]. But again, it’s a cluster trial, and I think it’s important to underscore also the difference between doing cluster and individual randomization because despite the fact that they are randomized after the current practice that we use in cluster trial, we often see, just like in the Gravenstein trial, that there sometimes are imbalances in between the clusters [20]. So, despite doing a very big effort in getting randomized data, we do sometimes, in these large-scale cluster trials, see that there’s imbalance [19]. And DANFLU-1 and DANFLU-2 trials that we just discussed earlier, they are in effort in trying to do individually randomized trials in very large scale [18]. So, both cluster and individually randomized trials within a real-world setting are currently ongoing with these different trials testing which vaccines we potentially should use for different patient populations or different populations in general [18,19,20].

Ann: Well, great. Thank you for that explanation.

Considerations for Data Evaluation in Vulnerable Populations

Ann: So now moving on, we’ll consider what are some of the elements that we have to consider when dealing with very vulnerable populations receiving flu vaccine? And I’ll just get the discussion a little bit started. When we think about enrolling older people with co-morbidities, either in randomized settings or in real-world settings, one of the issues that we face is that it’s very difficult to enroll very fragile people and people who have severe comorbidities. And therefore, we have to make sure in our trials that we build in stratification or numbers that need to be included so that those populations will be included. So unfortunately, the bulk of the data for the really frail population is immunobridging and effectiveness studies. And they are not without value, but we would be better served to make a better effort to get them into the trials to begin with. So, Stefania, maybe you could tell us about some of the key studies comparing high-dose to standard dose quadrivalent influenza vaccinse that have been done in some of these vulnerable populations.

Stefania: Yes indeed. And I think that actually for the high-dose vaccine, we do have quite a large body of evidence. And referring to what you were mentioning, I really like to mention the study by Gravenstein that has been carried out in more than 92,000 nursing home residents, and the high-dose has shown an efficacy of decreasing the hospitalization for respiratory illness by 13%, for example, and 21% decrease for pneumonia in this very vulnerable population as the nursing home residents [20]. And overall, I think that in terms of high-dose vaccine, I would simply mention the most recent literature review that includes six randomized clinical studies and 15 observational studies that have been conducted in 12 seasons on more than 40 million people, and where we have seen consistently more efficacy of the high-dose versus the standard dose in reducing all the clinical outcomes that we have mentioned before, such as not only the infection rate, but also the hospitalization for all causes– cardiorespiratory, cardiovascular, and pneumonia– in older adults [17, 21, 22]. Also, in the subgroups with comorbidity, with frailty. And this was consistent across the study seasons, and the study populations. And I think that therefore, the consistency of the results in both randomized clinical trials plus the observational studies and across the different populations give strength to the results of the higher efficacy of the high-dose versus the standard dose [17, 21, 22].

Ann: Well, thank you.

Concluding Remarks

Ann: So, this has been a really great discussion of what I consider a very important topic on how we get the best data from flu vaccination studies we can. I think that we’ve certainly seen that flu causes a very significant burden of disease, both direct respiratory illnesses, but all the indirect effects that we began our discussion with [1,2,3]. And quantifying the benefits of flu vaccine, as we’ve heard, can be very challenging, but certainly should include some of the following elements. And that would be a clinical outcome assessment that includes not only infection, but important outcomes such as medically attended disease and hospitalization [6, 12]. That we need randomized studies in both clinical and real-world settings that we heard a wonderful description about what’s being done in Denmark [15, 17, 18]. We can also do comparisons to standard of care. And while we have to be careful about intrinsic bias, there are ways to do good effectiveness studies where bias can be reduced [15]. And then importantly, we need consistency of results across all the seasons with study design and looking at multiple seasons to look at the different strains which circulate. And so far, the high-dose flu vaccine has been shown to have considerable effectiveness in older adults, and especially in some of these vulnerable populations [17, 21, 22]. So, thank you for listening.