In a commentary for Vaccine in January 2024, two familiar members of the vaccine community, Dr Gregory Poland and Dr Steven Black, explored the need for a “new vaccine safety surveillance paradigm” to encourage greater trust in vaccines. They reflect that, although vaccines are “generally safe” and usually provoke “only transient side effects”, there are a variety of other reactions that range from the “immediate and easily observable or measurable”, to the “clinically ‘silent’ or cryptic”. The latter are harder to identify and link to a vaccine.
“It is critical to be vigilant about rare, silent, or subtle reactions.”
Drs Poland and Black suggest that public health agencies and healthcare providers could play a “much more favourable and vital role” in securing vaccine trust by “enlarging” the current safety paradigm and, crucially, conveying these risks and benefits “transparently” to the public. While we explore the commentary in greater detail, we invite you to consider how greater transparency and detail could improve public trust.
Mistrust and hesitancy
The authors state that the late identification or recognition of “silent” or “cryptic” adverse events causes “public mistrust”, which is at “historically high levels” globally. Indeed, WHO recognised vaccine hesitancy as a “top ten” global health threat in 2019, and things have possibly worsened since then.
How can vaccine safety be better evaluated? The commentary considers how vaccine safety is “initially established” by randomised controlled trials (RCTs) and followed in post-introduction safety surveillance systems. The offered US example of the Vaccine Adverse Event Reporting System (VAERS) is a “passive surveillance” system that collects and analyses reports of adverse events after vaccination, yet these systems are “heavily dependent” upon observation of the adverse event.
Drs Poland and Black suggest that, “almost exclusively, adverse events detected by such surveillance systems are confined to those that fit one or more criteria”:
- Easily identifiable and result in medical utilisation
- Can be seen on imaging
- Have an associated biomarker
- Can be seen and modified by treatment
“As a result, adverse events that do not result in medical utilisation or cannot be clinically observed, measured, or imaged are, for the most part, ignored, remain potentially unidentified, and are of less interest or priority to public health agencies.”
Events on a spectrum
We are presented with specific adverse events and a “spectrum of events”. For example, Guillain Barré Syndrome (GBS), is observable, detectable, and measurable. The “pre-specified consensus definition” for vaccine associated GBS “depends primarily on an epidemiological definition and temporality”.
However, what about other associations, such as neuropathy, hearing loss, or tinnitus? The latter cannot be measured but “may substantially decrement quality of life and result in life-long disability”. Another example that the paper offers is the safety risk of aluminium exposure, a risk “only recently identified after more than fifty years” of alum use in vaccines. What about “side effects that uncover subclinical underlying pre-existing pathology?”, ask the authors.
“The point is that it is possible for vaccines to result in a spectrum of adverse events from asymptomatic, subclinical, mild, moderate, severe, or life-threatening.”
There is also a “continuum” of being observable, having a biomarker, or having an imaging component, that contrasts with those that are less readily detected, observed, imaged, or don’t have an associated biomarker.
“Currently, vaccine safety surveillance mechanisms have no interest and have made no proposals to identify such associations – despite the positive value this would have on public trust of vaccines.”
What’s in a name?
“Naming a phenomenon or new adverse event category requires a balance between being descriptive and succinct.”
The authors suggest the term Cryptic Adverse Event (CAE) for vaccine-associated side effects that are “silent or currently unobservable”, with implications of something that is “hidden or not immediately apparent”. The term also allows for re-categorisation when an event is more widely recognised or can be observed and classified with new technology.
Causality
Drs Poland and Black recognise that a “primary” challenge with these events is establishing causality. They consider the importance of determining whether the event would have happened in the absence of vaccination. Although RCTs are the “gold standard”, for rare and cryptic events they are “almost certain” to fail to identify them. Therefore, deliberate studies are required.
Post-market surveillance and Phase IV studies “have been critical” to the identification of possible causal events, but this is dependent on sensitivity to “readily diagnosable and reported adverse events”. Therefore, studies and systems “of the future” should “include mechanisms” to detect these cryptic but significant events.
“This likely requires active (rather than passive) reporting and solicitation of adverse events.”
For example, the authors consider an application of AI technology: screening of social media posts for new symptom complexes or concerns. They reflect that none of the Phase III COVID-19 vaccine trials detected or reported COVID-19 vaccine-associated menstrual disorders or irregularities.
“Had investigators actively solicited information on such issues or symptomatology it would have been detected and quantifiable.”
It’s noted that the lack of reporting by subjects could be a result of “not correlating an unexpected symptom with vaccination”, but “discussion of this specific issue occurred early in social media”.
What are some of the challenges that arise in detecting and verifying cryptic adverse events? The commentary offers subjectivity first; symptoms like fatigue or menstrual irregularities are “inherently subjective”, which presents a challenge in determining severity and onset of the symptom in relation to vaccine. Other challenges include pre-existing conditions or psychosomatic factors. However, “sensitive and high-quality” studies and systems can be conceived to “detect, at a minimum, differential effects between vaccine and placebo arms”.
The authors comment that safety and surveillance systems, in the US at least, are “excellent” at identifying adverse events that can be observed with biomarker and/or imaging confirmation. For events that are distant from vaccination or do not prompt patient reporting or medical utilisation, they are “far less good”. The delay in identification that follows allows public distrust to arise.
“We advocate that active surveillance for known and important but unknown symptoms or AEs be undertaken using not only electronic data collection after immunisation of lists of symptoms (much as physicians perform and collect a thorough systems review) but also through scanning of social media to identify new symptom complexes.”
The deployment of AI could start in Phase III studies and continue in post-licensure studies.
“It should be appreciated that high quality, exhaustive safety surveillance – transparently reported and followed up with additional research is the surest way in which to build public trust regarding vaccine safety.”
Predicting “pushback”, Drs Poland and Black state that “high levels of public mistrust” need to be addressed, and that much of it “could be answered by improved safety surveillance and transparent reporting”.
“It is essential to address concerns about potential adverse events comprehensively and compassionately.”
The final message in the commentary is of obligation:
“This we owe to those we immunise.”
Do you agree with the argument that Dr Poland and Dr Black present in their commentary, and how might their ideas be realised? Vaccine safety returns as a key track at the Congress in Washington, so do join us to continue this discussion there, and to hear from Dr Poland! For more updates from thought leaders, don’t forget to subscribe here.
