I am writing in response to the correspondence of El-Gizawy et al.1 who expressed surprise that aromatase inhibitors were deemed to have a sufficient evidence base for their use to be recommended. The recent systematic review of the effects of aromatase inhibitors on pain associated with endometriosis2 and my subsequent commentary3 make the point clearly that the data on the efficacy of aromatase inhibitors in alleviating pain associated with endometriosis are sparse and that robust randomised controlled trials are required. Considerably more data should be gathered on outcomes such as pain scores, quality of life and adverse events. However, I believe the initial findings have been sufficiently encouraging to recommend appropriately monitored usage of these agents. This usage should be limited to the small number of severe cases of women with pelvic endometriosis causing intractable pain who have not responded to other interventions such as nonsteroidal anti-inflammatory drugs, the levonorgestrel intrauterine system and gonadotrophin-releasing hormone analogues. El-Gizawy et al. also raise concerns about cost-effectiveness; in my opinion this is unlikely to be an issue as the use of aromatase inhibitors for this indication would only be for a very small group of women as third or fourth line therapy. Long-term therapy should of course be monitored with bone density scans and either cessation of therapy or low-dose add-back hormone replacement therapy (HRT) considered if significant loss of bone density occurs. There is currently very little, other than radical pelvic surgery with its potential risks, that women with intractable endometriosis-related pelvic pain can be offered. Under these circumstances, it is justifiable that aromatase inhibitor therapy, albeit with a limited evidence base, should be attempted.