It is quite evident from the reports we see but sometimes we fail to recognize it, that the confirmed victims of bird flu to date have been overwhelmingly young compared to what we see in seasonal flu. Here is the age distribution in the latest WHO tally, as graphed for us by WHO's Pacific Regional Office:
Source: WHO Western Office for the Regional Pacific
There are two principal reasons I can think of for this difference in pattern, and my guess is both are operating simultaneously. The first has to do with what epidemiologists call observation bias. Death from acute respiratory failure in the young is more unusual than in the elderly so they get better diagnostic work-ups that result in confirmed cases. The claimed association with poultry works in the same way. It is the young and young adult most likely to have close contact with infected birds, and since this contact is a feature in their clinical history that calls attention to the possibility of H5N1 infection, they are diagnosed more readily than the elderly who don't give this history. That leaves open the possibility that other kinds of exposure might be involved, something we see no reason to rule out with the present evidence.
Bias, in epidemiologic lingo, is systematic error, i.e., the skewing of the disease to the young by observer bias is thus more apparent than real. But the other possibility is that the skewing is indeed real. A hallmark of influenza pandemic years is a shift of the age distribution toward the younger age groups. In 1918 the mortality shifted from the more usual U-shape (infants make up the left side of the U, even in non-pandemic years) to the notorious W-shape, the big hump in the middle being young adults who were hit disproportionately hard. No one knows exactly why there is this W-shape. Why, for example the valley in the 5 - 15 year old group?
One of the characteristics of the current H5N1 cases, however, is that the 8 - 12 year old "safe" ages don't seem to be safe any longer. The histogram shows quite a bit of mortality and morbidity from confirmed H5N1 in the under 20 age group. Whether this reflects some innate susceptibility to the virus in this age group, reflects the pattern of exposure, or is just another reporting artifact we don't know as yet. We'll take the safe way out and say it is probably a combination of all three. We realize this isn't very informative without estimating the relative importance of these factors, which, unfortunately we have no way of doing and about which we have no opinion.
Addendum: Reader William Milliken, Missiontech.org, was kind enough to send along this redisplay, by decade and separating deceased cases, of the same data:
- Log in to post comments
As others have previously commented, everything is in decades, except for the first two age groups.
It would have given more information to a glancing public - if governments wanted public attention, which they don't-
to have the first bar to be Birth to age 9.
The data reports only number of cases, not percentages for each age group.
The countries where H5N1 have appeared have a population pyramid with a broad base. This means they have many more people in the younger age groups than in the older generations. Hence if there was the same percentage of infection in each age group, the graph of cases and deaths might look much like the above at the elderly end, but have more cases in the under 10-year olds.
I haven't yet seen cases and deaths as a percentage of the people in the various age groups.
rs: The comparison is with age specific ratios for seasonal flu, which has the same denominators. Thus we can compare distributions.
The figures I've seen for 1918 Spanish flu have the 10 to 19 age group as the least affected, with the 50 to 59 group second least effected.
The most effected were children under one year, where more than 3 percent of the cohort dies. The second most effected were the two groups in the 20 through 39 range.
If you graph the deaths from the 1918 flu, based on age, the graph looks like a slightly skewed letter "W". I wonder if H5N1, if and when it goes pandemic, will affect the population the same way, or if it will maintain its current look. My bet is on the 1918 look, but either way it's going to be "interesting", as in the old Chinese curse, "May you live in interesting times."
-- Lenn
I see here two categories, a, affected (->39), b, rare, with a potential cutting off point that might be 35. I'd formulate: "Those born after 1967 are .." - a different way of looking at it.
To believe in a SLOPE in the first block (0-39) I'd want more. It will probably turn out to be a flattish line with a trailing off, a seguing into the block 2 (older group) picture. ??
I'm doubtful about the contact aspect. Within the explanation, contact must be taken to mean close contact (otherwise, no age difference, as a family living 300 meters from infected birds would all be exposed similarly), such as playing with infected birds, caring for them, taking them to market, killing and plucking them, cooking them, etc. I can't see that as applying to block 1 vs. block 2.
There is no social / activity classification that does the job. In fact, if such classes were playing a role, one would expect differentiation (not similarity) in the young group, which breaks up into babies-toddlers/roaming children/teens/young (working...) adults. If close contact was a factor, naively (as I don't know squat about avian husbandry and how things go in China for ex.) I would expect women, and particularly elderly women (home bodies, cooks, take care of the coops, etc.) and perhaps children 4-13 to be more affected, but even that seems stretching it a bit.
Nah, none of that works. So are women more affected than men? I tried to look it up with no success.
Also, one would have thought that if habitual or frequent close contact played a role, by now the questionnaire data and numbers would have thrown up that link? I mean, on a 'contact scale', there is no comparison between those who own and keep chickens, those who sell birds, those who work in the food industry, those who live in a rural setting with all kinds of animals about, etc. etc. and those who do not. And that division would never be directly related to age, except through activity (e.g. industrial food worker vs. child who does not work.)
Some sort of Cytokine Storm type phenomenon?
revere, I personally think the distribution is absolutely real. I want to thank william for sending the alternate chart, as the WHO chart is misleading by splitting the under-10 into 2 bars, thus giving the impression of a lower than actual incidence.
In addition, I don't know where this theory of the young being more exposed to poultry comes from. Certainly in SE Asian village life, chickens just run around underfoot the whole time, so pretty much everyone gets exposed more or less to the same degree, IMO.
Westerners might have the idea that elderly people do less active 'work' and therefore get exposed less. This is far from the case.
First of all, notice the sharp drop in incidence after age 40, which is hardly 'old'. Secondly, the elderly tend to help out with easy chores, and feeding chickens would be one of them, as compared to working in the fields. Thirdly, elderly people are the major source of childcare. They often sit around outside their huts on low stools to watch the little ones as they play in the dirt. In that sense, they get just as much exposure as the children, IMO.
This notion is very much an example of using an idea to fit the data, and very few people have publicly challenged it so far.
anon_22: As I understand the exposure idea, it is that it is the special chore of the very young to get the eggs in the henhouse in the morning. Which chart you prefer depends on what you want to see. The age categories in the WHO chart are conventional and recognize that there is a difference between 0-5 and 5-10. It also provides more information.
You might be right that it is real, but I don't know how we can know this at the current state of the evidence.
The actual care that these people got for the dead and survivors is also not taken into account. Plenty of ventilators for 200 plus cases as a rule to go around in the world. Plenty of care as well. Those numbers against 1918 flattened out considerably at least for the cases on the graphs that I have seen. Seems that everyone got sick, and most recovered on their own. But that was a 5% bug. The overall on this right now is about 67% and higher than that for this year at least in Indonesia for CFR. Genetic predisposition is also something thats being kicked about. So where does this fit into all of that?
revere,
I'm writing this from first hand experience, albeit from many years ago. Where I grew up there were plenty of neighbours who kept a few chickens, and you can find them wherever people are. We used to have to be careful not to run barefoot cos of the chicken droppings that you can slip up on!
And when people take their meals, the chickens would be under the table between people's legs getting the scraps that fall off.
On the matter of getting the eggs, again both from first hand experience and from my understanding of Asian life even now, I've never come across that as a custom at all. If anything, the eggs are the precious commodity, not to be trusted to kids. Their chore would more likely be feeding them or rounding them up at the end of the day.
You're the epidemiologists but my observation of village life tells me that pretty much everyone gets exposed, cos of the chickens running around picking at the dirt. Remember that photo of the survivor from the Karo cluster in 'hospital'? He was sitting at an open doorway and right outside just beyond the threshold you can see a chicken picking up food on the ground!
It might seem a trivial point, but this illustrates the importance of understanding how people live. Now I could be wrong, but I would like to see some definitive evidence to be convinced.
To me the next obvious question is the percentage survival for various age groups. It does look like the survival rates for the youngest groups is better... approaching 50% for the under 5 patients... as opposed to ~35% for those aged 20-40 and ~25% for those 10-19.
But while it's been too long since I took statistics for me to do the math, I'm pretty sure that with this small a sample size, even if patients received indentical care, there'd be a large margin of error. And since we know they didn't, we don't even know for sure that there's an real difference in the odds of survival among a 5-year-old, a 15-year old and a 35-year-old. Or if having lived through the last pandemic provides any protection.
No wonder public health authorities are having trouble coming up with treatment/vaccination priority lists.
Missing data which would be helpful to evaluate this chart is how many of those victims of the flu received tamiflu, and how long after the onset of symptoms was it administered in survivors as compared to those who succumbed. I'm presuming most survivors were on tamiflu and ventilators.
Regardless of what the apparent cfr in the graphs are, if a pandemic strikes we will rapidly be unable to offer tamiflu and ventilators to more than a small fraction of the infected population. (which I believe is what MRK was talking about in post above). Thus IMO we would see a much lower survival rate, despite our advantage in terms of "modern medicine"...which can't do much good if it can't be provided. It truly is a frightening thought to imagine yourself or your loved ones choking to death on their own body fluids and knowing that there is no 911, no hospital to rush to, no ambulance to call, no doctor or medicine to help.
My point entirely MIH. They are asking for 1.5 billion more for bird flu and 750 million a year thereafter for really nothing more than a compensation program. Want to get bird flu in the US. Tyson, Pilgrims Pride etc. will reap billions of dollars for nothing more than an open ended paycheck. All they will have to do is get an infected bird in.
Also all of the extreme measures that would be taken to "care" for someone would be very, very uncertain. That is to say that we could spend BILLIONS for care that is nothing more than a palliative. Hospice for the lack of a better term. Nothing has ever been indicated as a sure fire way to get a survivor. Yesterday the Canucks put up a controversial "triage" plan that really is nothing more than a methodology on cullling people for the few supplies that will be there. Mind that in the US there are only about 2 weeks to a month worth of medical supplies. After they are gone, thats it...Why? They are made in SE Asia and India.
What about some kind of immunity from prior exposure to h2n2? It stopped circulating in 1968?
Just noticed the dates match. I learned everything I know about the science of flu on this site. I know, I need to go back and reread it again....
Steph