1. Summary of quality assessment of included studies.
| Study id | Assignment | Alloc concealment | Baseline | ITT | Dropouts | Outcome Blind |
| Chang 2002 | Yes | Probably Adequate | No detail presented | Yes | Yes (7/120) | Yes |
| Hartkamp 2004 | Unclear | Adequate | No (postmenopausal status higher in active arm; oestrogen use lower in active arm) | No (2 declined final dexa scan excluded | Yes (2/60) | Yes |
| Nordmark 2005 | Unclear | Unclear | Similar (DHEA treatment group had slightly more active disease | Unclear | Yes (3/41) | Yes |
| Petri 2002 | Unclear | Probably Adequate | Yes | Yes | No (49/191) | Yes |
| Petri 2004 | yes | Adequate | yes (although Anti dsDNA higher in active treatment group) | Yes | Yes (115/381) | Yes |
| Van Vollenhoven 1995 | Unclear | Unclear | Yes | No | Yes (2/28) | Yes |
| Van Vollenhoven 1999 | Yes | Probably Adequate | No Males in placebo group only, organs effected differed.Some differences in baseline activity scores with DHEA group tending to be higher (none statistically significant) | No | Yes (2/21) | Yes |
| KEY | Assignment random | Allocation concealed | Baseline characteristics similar | Intention to treat analysis | drop outs described | Outcomes blinded |