Interesting. Almost definitely something to do with adenoviruses then. I would imagine that Sputnik probably also has this side-effect. I wonder if "wild" adenovirus infection does this too.
Deciding whom to give these vaccines to is a complicated decision for each country to make and depends on:
a) The incidence rate for each demographic of CSVT-type clotting from the vaccine[s].
b) The expected distribution of clinical outcomes from those events, making reasonable assumptions about likelihood of early detection and outcomes with treatment for heparin induced thrombosis.
Those two give you the clinical cost.
Then you compare those with the clinical benefit from:
1) The relative likelihood of infection for a person, given their demographics, personal exposure profile etc.
2) The absolute likelihood of infection in the country / region in question based on assumptions about the future of the pandemic
3) The expected distribution of clinical outcomes from infection given the person's characteristics
(2) is important because you're comparing a one-time risk of side-effects with a time-related risk of infection.
Clearly, the same person in Manaus or Adelaide is not really making the same trade-off even if in the abstract, the side-effects and IFR are the same.
You also need the counterfactual: If you decide not to give an ad vectored vaccine, you are not deciding never to vaccinate, you are deciding that this person will get a different vaccine, probably later. How much later depends on the location. If you're in the US, you've got Moderna and Pfizer/Biontech coming out of your ears so it might only delay any individual vaccination by a few weeks. If you're a healthy 25 year old who can work from home, the risk trade-off of having to wait an extra 6 days is very different than that same person if they have to wait another 6 months.
As a result, different countries, looking at the same data will rationally make different decisions. The US has lots of mRNA vaccines so will choose one thing, the EU has been badly hit by the failure of Sanofi/GSK to deliver a working vaccine (and by AstraZeneca scaling problems) but has quite a lot Pfizer production capacity, the UK has used up its Pfizer stocks and will now only use it for second shots. CureVac and Novavax may be able to supply some doses at some point but aren't in the picture yet. Several European countries are negotiating to get Sputnik but that is also ad vectored so may have the same problem. That changes the risk calculation because it changes the time that people will remain unvaccinated in the counterfactual where ad vectored vaccines are not used for their age group.
All of this is before the complicated public health element of keeping messages for the public simple and the issues caused by changing that message frequently.
I don't know to what degree countries that had set relatively high minimum ages for the use of AZ had been counting on J&J for younger people. In many cases, AZ deliveries were sufficiently slow due to yield problems that it was "free" from a scheduling point of view to restrict it to only the oldest since the expectation was that they would only receive enough for those anyway and could use other vaccines for younger groups. If J&J was a big part of that, then they now have to update that calculation because the counterfactual to using AZ for those people is no longer "do them with J&J next week". That may be hard to communicate to people who had been told that a particular vaccine was not safe in their age group previously but may now be asked to take it anyway.
I'm glad that I don't have to make these decisions. They are not easy ones.