After multiple imputation, the failure event number has changed. How can I solve this problem?

[Byounghwi Ko]

Hello stata users,
I have a problem for multiple imputation...



I was willing to impute BMI, Hemoglobin, albumin, CRP, LDL, ejection fraction by pmm and ALC1, LVMI by logistic regression.
After that, I did mi stset for DEATH_YEAR and DEATH_EVENT to try an "Univariate" Cox regression by Intensity_3.

And this is the result.

When I do the univariate cox regression, the number of failure in the dataset changes.
Did I do something wrong code for multiple imputation?

Thank you for reading....

[Byounghwi Ko]

I solve it!