How to assign NA as the value using mutate() and case_when() in R

Question

I have a dataframe column consisting of numbers (e.g. with unique values of 1, 2, and 99).

library(dplyr)
df <- data.frame(A = c(1,2,1,1,99,2,1,99,1))

I want to change all rows where the value is 99 so that the value is NA instead. I tried doing this with mutate() and case_when():

df <- mutate(df, A = case_when(A==99 ~ NA))

...but it converts all rows to NA (rather than just the rows where the value is 99).

Previous Stackoverflow questions similar to this (e.g. this one) state that using NA_double_, NA_integer_, NA_real_ etc fixes this. I tried those but this either:

(a) changes all values in the dummy example I've given here, or (b) I get an error on my real data (which is a haven_labelled variable):

Error in 'mutate()':
In argument: 'A = case_when(...)'.
Caused by error in 'case_when()':
! Failed to evaluate the right-hand side of formula 1.
Caused by error:
! object 'NA_double_' not found

I know I can set the correct values, and only the correct values, successfully using the method below:

df$A[df$A==99] <- NA

... but I'd like to understand how to do this using mutate() and case_when().

What should I change my code to so that it works using mutate() and case_when(), whether it's (a) the dummy data presented in this question or (b) the haven_labelled data in my project?

Note: This is not a duplicate of this question because I am asking specifically about using case_when() with mutate().