Lenalidomide to control gastrointestinal bleeding in hereditary haemorrhagic telangiectasia: potential implications for angiodysplasias?

Hereditary Haemorrhagic Telangiectasia (HHT) is an inherited disorder of abnormal blood vessel formation. It typically causes mucous membrane bleeding that progresses in severity with age (Begbie et al, 2003). Current systemic therapies are unsatisfactory, as are local measures because of the widespread nature of the condition. Thalidomide is an antiangiogenic drug that anecdotally controls severe gastrointestinal bleeding in a range of angiodysplasias, including HHT (Bauditz & Lochs, 2007). Unfortunately, thalidomide has many side effects. Lenalidomide, which is better tolerated than thalidomide in myeloma (Dimopoulos et al, 2007), is also antiangiogenic. We report a patient whose life-theatening bleeding was controlled with thalidomide but was better controlled on lenalidomide, and with minimal toxicity. This is the first reported use of lenalidomide to control bleeding in an angiodysplasia. A 69-year-old female with HHT suffered chronic gastrointestinal bleeding as well as sudden massive life-threatening bleeds. She started thalidomide 50 mg, increasing to 100 mg, daily. Her telangiectasia diminished, and all visible bleeding and transfusions ceased except once when the dose had been reduced to 50 mg alternate days because of side effects. Bleeding disappeared again on 50 or 100 mg daily. Eventually neuropathy developed as a side effect of the thalidomide and the drug was stopped. Lenalidomide 10 mg daily was started, increasing to 15 mg daily. After 15 weeks on 15 mg daily she became pancytopenic, with Hb 104 g/l, white blood cell count (WBC) 1·3 × 109/l, neutrophils 0·6 × 109/l and platelet count 56 × 109/l. The dose was changed to 15 mg alternate days and her blood counts recovered. The lenalidomide was then increased to 25 mg alternate days and she has remained on this dose for 7 months with WBC c. 2·5 × 109/l, neutrophils c. 1·5 × 109/l and platelet count c. 100 × 109/l. She has remained free of gastrointestinal bleeding and drug symptoms, with a higher haemoglobin and reduced iron requirement than during thalidomide or before treatment (Table I). Our patient’s freedom from visible bleeding, improved mean haemoglobin, decreased iron requirement and decreased telangiectasia on thalidomide and lenalidomide suggest that these antiangiogenic drugs prevent bleeding by inhibiting vessel formation. Our data suggest lenalidomide may be more effective than thalidomide. This concurs with in vitro data that lenalidomide is a more powerful antiangiogenic drug than thalidomide (Dredge et al, 2002). Lenalidomide probably acts partly by preventing endothelial cell migration (Dredge et al, 2002). Although the activity of these agents is most likely to be due to an effect on endothelial cells and vascularization, many other activities (some unexpected) have been described for both drugs. These include tumour necrosis factor α inhibition as well as complex immunodulatory effects. Recently, lenalidomide has been shown to inhibit regulatory T cell suppressor activity, a property that is not shared with thalidomide (Galustian & Dalgleish, 2009). It is therefore possible that more than one property or susceptible pathway is required for these agents to be effective in this condition and that lenalidomide is more effective than thalidomide by virtue of multiple actions. Alternatively lenalidomide may be more effective on a singular pathway not fully elucidated. A hangover effect from the thalidomide treatment period is unlikely because our patient’s haemoglobin increased and iron requirement decreased whilst taking lenalidomide, suggesting the drug was active. Daily lenalidomide dosing was limited by marrow suppression but this was overcome by alternate day dosing. The appropriate dosing schedule is not known but alternate day dosing is not unreasonable based on the rate of vessel regrowth in the rat model (A. Dalgleish, personal communication). The antiangiogenic activity of thalidomide and lenalidomide is probably dose-dependent (Dredge et al, 2005). Our patient’s bleeding on a reduced dose of thalidomide, which disappeared on increased dose, supports this dose-dependence and has been noted in a previous report (Heidt et al, 2006). Because of this dose dependence, the Lenalidomide dose has been kept as high as possible, within the limits of marrow suppression. Our patient’s haemoglobin may have been even higher if she did not suffer marrow suppression. Lower doses might be sufficient with no marrow toxicity. Thromboembolism is a risk with lenalidomide treatment in myeloma (Dimopoulos et al, 2007) but whether it is a risk in HHT is unknown. Thromboprophylaxis may be appropriate despite the bleeding risks. Our data suggest that lenalidomide is more effective than thalidomide in controlling gastrointestinal bleeding and is better tolerated. Given that thalidomide appears to be useful in controlling gastrointestinal bleeding in many disorders characterised by abnormal blood vessel formation (Bauditz & Lochs, 2007), our report may have significant implications for use of lenalidomide in the management of a range of angiodysplasias. We would like to thank Prof A. Dalgleish for his helpful comments.