Mass loss via rotation speed increase.
Depicted: V838 Monocerotis. This star emitted a bright flash while Hubble was watching. The flash allowed Hubble to see the cloud of dust surrounding it. How would such a cloud form? I will propose a mechanism that is my answer to this question.
Backing off a step: how could a star eject material that might form a dust cloud? Some stars like Wolf-Rayet stars do this fairly often. One mechanism would be an acceleration in the speed of spin. Just as speeding up the spin of a merry-go-round with a motorcycle (check youtube if curious) will cause riders to be flung off of it, so speeding the rotation of a star will cause it to fling mass off of it.
The lost mass could expand to form a cloud or ring surrounding the star.
How to increase rate of spin? One way would be to decrease the size of the star. As a skaters spin will increase when she pulls in her arms and legs, so a star's rate of spin will increase if it decreases in size. The aforementioned Wolf-Rayet stars have shrunk down as compared to their prior size, and maybe consequent acceleration of spin accounts for their dust generation?
I think these Wolf-Rayet stars are older than middle age and so not what is requested in the OP.
Another way to increase rotational speed is to acquire it from another object. If I go running at a merry-go round and jump on to the edge, it starts turning - it acquires rotational momentum from me. Some people think that V838 Monocerotis was originally a binary star and the observed flash was the collision of the partners.
Tylenda disagrees: "I think that almost all of the red novae are
mergers." In particular, he argues that the best-known red nova, V838
Monocerotis, resulted from such a merger. That explosion was more
powerful than V1309 Scorpii, indicating a greater mass.
For contact or semi-detached binaries, the transfer of mass from a
star to its companion can also result in a significant transfer of
angular momentum. The accreting companion can spin up to the point
where it reaches its critical rotation rate and begins losing mass
along the equator.
If prior to the collision one of the partners had its rotation accelerated by the other, that could account for ejection of the dust clouds which the later collision illuminated.
How then to make an increase in acceleration for the middle aged star described in the OP?
1: Star is actually a binary, and one partner is gradually being accelerated by the other. When acceleration of partner exceeds critical rotation rate, it loses mass to form the dust cloud.
2: Star is obliquely impacted by an outsider arriving from interstellar space - rogue white dwarf or what have you. The mass transfer from the object is not much, but much of the kinetic energy transfer turns into faster rotation of the star, pushing it past critical rotation rate and causing mass loss into the dust cloud.
3: Star's mass is increased by an outside impactor, with consequent decrease in diameter and increase in spin rate.
After my google session I am aware that there are many schools of thought about V838 Monocerotis and the provenance of its cloud, and Wolf-Rayet stars, etc. I do not mean the above to represent dogma on these entities. I did want to invoke real things to illustrate the idea that faster spin of a star could cause it to emit a dust cloud.