Owen Yang
I actually want to hear more about immune age, and I am fascinated by it. But a lot of time when I hear it all I feel is disappointment and sorrow.
Immune age is such a brilliant idea
We all want to know whether our immune system is older or younger for our age. If we are immunologically older, we might be at higher risk of infections, and we might want to do something about it. We might want to take some measures to ‘booster’ our immunity, or simply to take extra caution to prevent infection.
The idea of an immune age is also consistent with a realistic expectation. Instead of trying to stay young, we accept that ageing itself is a natural process, and therefore we set the goal at a realistic range appropriate to our age.
Age is time since birth
Unfortunately, there are many loopholes in the development of this concept. More work need to be done logically, sociologically, and mathematically before we run into a solution to achieve this concept. For example, age is time since birth, and what does it to do with anything we care? If anything, age is in its first half an indicator of growth and development, but in its second half an indicator of time to death. The longer we live, the shorter we have left. Because it is not straight-forward to measure time to death, the best obvious indicator is to use time since birth to predict time to death.
I do agree death seems to be such a over-simplified outcome here, and should be broadened into a more sociological concept. Instead of predicted risk of death, in a medical scenario we could be thinking about risk of loss to function, risk of cardiovascular disease, risk of infection or risk of loss to immune function (some sort of ‘immune failure’). We care about age because loss of function is more likely or more imminent when we live longer, but in fact it is the time we have left for ‘adequate’ function that we care at the second half of life.
An ideal way to measure age in the second half of life
Therefore, an ideal way to measure ‘age’ in the second half of life is not time since birth, but time to death, time to loss to function, or time to all age consequences that we care about as an individual or as a society. This time-to-doomsday is inevitably correlated to time since birth (the chronological age) at a larger time scale (for example when you compare age 50 to age 80), but should be better correlated to the doomsday at a smaller time scale (for example in the age range of 60-64). This is making sense if you think about it: heart function should be correlated with age (time since birth) better than with time to death in a large range of age 50-80, but could be correlated with death better than age in a small range of age 60-64. You may not say a 64 year old will die sooner than a 63 year old, but you may guess so if the 64 year old has much poorer heart function.
There is an overall assumption that our body is ageing as a whole, but different aspects of our body may age at different speeds. Therefore, in addition to an overall ageing age, we would also like to know our cardiovascular age, immune age, telomere age, so that we know what our weakest link is. We can then prioritise restoring the function or prevent rapid decline of the function of that aspect.
Even so, we should not be confused with what we actually care, that is time-to-doomsday, not time since birth. The time to doomsday is a function of risk of doomsday. For example, if the doomsday is death, then it is a function of risk of death. If the doomsday is heart attack, then it is a function of risk of having heart attack. It should be relatively easy to imagine: if the risk of heart attack is 50% every year, then the median time to heart attack is approximately 1 year. If the risk of heart attack is 10% every year, then the median time to heart attack is approximately 7 years. The actual calculation can be slightly more complex but I will keep it simple here. The median time depends on the risk.
Therefore, what we really need to find are the indicators that predict the doomsday better than time since birth at this scale. For example, we might find hypertension, body mass index, smoking, or evidence of atherosclerosis predict heart attack better than age at this scale. Framingham score or QRISK score is then devised to predict this risk. You can compare your risk to some sort of healthy population norm by age and know whether your risk is better or worse than this norm.
Statistical practice of immune age that needs a second thought
What is difficult now for immune age is there is no clear definition of an immune failure, and immune is not accepted as an overwhelmingly major cause of death or any major events. It is possible that immune is a major underlying driving force of many things that we can observe, such as some sort of tissue-level inflammation that lead to cancer, cardiovascular event, or other health conditions, but the extent of this hypothetical driving force is not proved to be major. Therefore, it is currently unlikely that we can find immune indicators that can predict death reliably at a population level, and even if we could, there is not yet a clear clinical, biological, or molecular pathway of this prediction. The stage we are at is probably still observing the change of immune profiles as we get older, and even so the data is very limited. A recent paper in Nature Medicine published some precious longitudinal data, and for those who are capable you could find it here.
Some people have started to get creative. Instead of predicting time to doomsday, they used all sorts of measured indicators to predict time since birth, and claimed this is the immune age. I have to admit I could not get my head around it to comment what is wrong with it. It is likely that the so-called immune age in this way is a modelled time since birth, and therefore can have similar effect of time since birth. However, I cannot expect this modelled time since birth to predict doomsday better than time since brith at any age scale. I need more time to think about this.
What do you think?
