Those bags are said to have been loaded in the rear hold which didn't help the CoG. A further 10 bags were left behind. The status of each bag is not immediately clear however the accepted number of bags loaded that were not on the load sheet is 19.
The document at the link below shows that right rudder was being applied progressively during the take off run. The key data is about 1/3 down the page showing larger lateral accelerations from 100kts onwards. This could be consistent with the hypothesis that the missing spacer was affecting performance prior to the tyre burst event and reliable witness statements that smoke was seen far earlier than the fire. I can't see any reference to the spacer in the report but I have not read it in great detail
https://www.bea.aero/docspa/2000/f-s...sc000725pa.htm