Michaelis-Menten parameters kept when going from mm-elimination to linear elimination

The following snippet will create code where the Michaelis-Menten parmeters are still in the model

```r
m <- pk_model() +
    dataset("mx19.csv") +
    pk_absorption_fo() +
    pk_distribution_1cmp() +
    pk_elimination_nl() +
    obs_additive(~log(C["central"]), var_add = 1)
render(m)

m <- pk_elimination_linear()

render(m)
```
