I think they are saying that, as use of cotrimoxazole is not recorded in the registry, they don't know which of the patients also has TB (which is treated with cotrimoxazole, hence treatment with this drug can be used as a marker for TB diagnosis), and so they can't identify which patients are suffering both HIV and TB. The group of patients that is being treated for simultaneous HIV and TB would be the "HIV+TB treatment cohort", i.e. not necessarily anything to do with an actual 'cohort study'.

It's hard to be sure without knowing what type of thing YYY and XXX stand for (knowledge of status, a drug, inclusion on a register...), but this would be my suggestion without further info. Hope this helps!

--------------------------------------------------
Note added at 1 hr (2015-10-27 14:36:02 GMT)
--------------------------------------------------

Ah ok, that makes sense. Your wording is almost fine, but I would include a "therefore", as it's the lack of inclusion of data on cotrimoxazole that is causing the lack of availability of a TB-HIV treatment cohort.

With a little tweak:

"Data on cotrimoxazole coverage are not currently included due to weaknesses in the registry, therefore there is no TB-HIV co-infection treatment cohort available."

Note, the data on cot. might not be absent/poorly recorded, they might just be poorly coded or otherwise unhelpfully rendered within the register for some weakness in the register's structure or something, so should probably stick to a closer translation where it's the register at fault and not necessarily the actual data input process (i.e. recording). That's probably a pickier response than you wanted, sorry :)