Do the participant size decrease when you remove an extreme outlier when conducting an independent sample t test? or do you keep the participant size the same although the outlier was removed?

Example: I have 18 participants and one came back as an extreme outlier, will the participants remain 18 or go down to 17?