how can we tell dagitty in R to not include a latent variable in the set of all the adjustment sets?

Question

I am using dagitty package in R. I have a DAG as follows where U is latent:

library(dagitty)
dag <- dagitty("
dag {
U [latent]
U -> Z1
Z1 -> Z2
Z2 -> Y
U -> X
X -> Y
}
")

When I call the adjustmentSets(x = dag, exposure = "X", outcome = "Y" , type = "all") function, this is the results:

{U} {Z1} {U, Z1} {Z2} {U, Z2} {Z1, Z2} {U, Z1, Z2 }

the results include U. How can I not have U in the results without changing the dag?

I expect it to not show U in the results. I tried changing the DAG as follows:

dag <- dagitty("
dag {
Z1 -> Z2
Z2 -> Y
Z1 <-> X
X -> Y
}
")

And I get the following results which is what I expected:

{Z1 } {Z2 } {Z1, Z2 }

But I don't want to manually change the results, as I have a really large graph and I want to get the right results without changing the graph structure.