I followed the guidance on sex standardisation from here https://www.stata.com/manuals13/rdstdize.pdf. It is my understanding that regardless of the method of calculation, comparisons of adjusted prevalence rates will be valid only if they are derived from the same standard population. For example, if one wants to compare age-adjusted rates for ischemic heart disease for males and females, then both the male and female rates must be standardized to the total (combined male and female) standard population. If the female rates were standardized to the female standard population and the male rates standardized with the male standard population, then the resulting standardized rates could not be compared across the male and female groups because the standard populations of each group are different.

In the case of the dstdize command I notice that when I code this:
dstdize stroke_any pop age_category, by(city) using(pop_weights_city)

*were stroke_any is the binary 1/0 population im analysing prevalence for, pop is my pop=1 and age_category is my age groups by city and using the population weights of the census in pop_weights_city.

The results that I get don't count the combined male and female population but rather just the males or females making my results not comparable. I did this manually and I see that difference. Any idea about this or how to sort this?