Recombinant human erythropoietins: new advice for prescribing
- Medicines and Healthcare products Regulatory Agency
- 1 December 2007
- Therapeutic area:
- Cancer and Urology and nephrology
Over-correction of haemoglobin concentration may increase the risk of death and serious cardiovascular events in patients with chronic kidney disease; it may increase the risk of thrombosis and related complications in patients with cancer.
Article date: December 2007
Recombinant human erythropoietins should not be given to patients with cancer who do not fulfil the criteria in the authorised cancer indications.
Recombinant human erythropoietins (r-HuEPOs) stimulate erythropoiesis. They are indicated for the treatment of anaemia in patients with chronic kidney disease. Some r-HuEPOs are also authorised for the treatment of patients with nonmyeloid cancer who develop anaemia after chemotherapy. Five r-HuEPOs are authorised in the UK: epoetin alfa (Eprex); darbepoetin alfa (Aranesp, a hyperglycosylated epoietin derivative); epoetin beta (NeoRecormon); epoetin delta (Dynepo▼); and methoxy polyethylene glycol-epoetin beta (Mircera▼). Biosimilar analogues of epoetin alfa have also been granted Marketing Authorisations.
The safety of r-HuEPOs has been reviewed because recently published data from clinical trials have shown a consistent, unexplained statistically significant excess mortality in patients with anaemia associated with cancer who have been treated with r-HuEPOs. Furthermore, results of studies suggest that treatment of anaemia with r-HuEPOs in patients with chronic kidney disease to achieve relatively high target haemoglobin concentrations may be associated with increased risk of mortality and cardiovascular morbidity.
Patients with cancer: risk of tumour progression and reduced overall survival
4 large controlled trials (including two unpublished studies) have assessed survival and tumour progression1 2 3 in a total of 2,833 patients (see table). 2 studies recruited patients who were receiving chemotherapy. Target haemoglobin concentration in 2 studies was more than 13 g/dL; in the remaining three studies it was 12–14 g/dL. The open-label study recorded no difference in overall survival between patients given r-HuEPOs and controls. In the 4 placebo-controlled studies, the hazard ratios for overall survival ranged from 1·25 to 2·47 in favour of controls. These studies have shown a consistent, unexplained statistically significant excess mortality in patients who have anaemia associated with various common cancers who received r-HuEPOs, compared with controls. Differences in overall survival in the trials could not be explained satisfactorily by differences in the incidence of thrombosis and related complications between the r-HuEPO groups and the control groups. The extent to which these outcomes might apply to the use of r-HuEPOs to achieve haemoglobin concentrations lower than 12 g/dL in patients with cancer who are receiving chemotherapy is unclear because few such patients were included in the data reviewed.
|Study (number of patients)||Treatment||Design||Cancer type||Hazard ratio (95% CI)|
|Overall survival||Tumour progression|
|Henke et al1||Epoetin beta||Double-blind, placebo-controlled||Squamous cancer of head and neck (patients given radiotherapy)||1·39 (1·05–1·84)||1·69 (1·16–2·47)|
|DAHANCA 10 (n=484)||Darbepoetin alfa||Open-label; control group received radiotherapy alone||Squamous cancer of head and neck (patients given radiotherapy)||No difference||1·1; p=0·01|
|20010103 (n=989)||Darbepoetin alfa||Double-blind, placebo-controlled||Various solid tumours (patients did not receive radiotherapy or chemotherapy||1·25 (1·04–1·51)||Not available|
|Leyland-Jones et al (n=939)||Epoetin alfa||Double-blind, placebo-controlled||Metastatic breast cancer (patients given radiotherapy or chemotherapy)||1·36 (1·05–1·75)||No difference|
|Wright et al (n=70)||Epoetin alfa||Double-blind, placebo-controlled||Non-small-cell lung cancer (patients given non-platinum chemotherapy)||2·47 (1·05–5·83)||Not available|
Table - summary of clinical trials of r-HuEPOs in cancer setting
A systematic review has analysed more than 9,000 patients with cancer from 57 clinical trials.4 Meta-analysis of overall survival showed a hazard ratio of 1·08 in favour of controls (95% CI 0·99–1·18; 42 trials, 8167 patients). Patients assigned r-HuEPOs had an increased relative risk of thromboembolic events compared with controls (1.67 [95% CI 1·35–2·06]; 35 trials, 6769 patients).
The available data does not enable accurate definition of a target range for haemoglobin concentration that has a consistently favourable balance of risks and benefits. However, no advantage has been shown for a haemoglobin concentration higher than 12 g/dL in patients with cancer. The purpose of r-HuEPO treatment is to relieve symptoms of anaemia and avoid the need for blood transfusion. Treatment should stop when symptoms of anaemia have been adequately controlled. Symptoms of anaemia may be controlled in some patients at haemoglobin concentrations lower than those conventionally considered to be normal.
There is currently no evidence to suggest that r-HuEPOs may adversely affect the risk of tumour progression and overall survival in patients with cancer who meet the criteria in the authorised cancer indications. The safety of r-HuEPOs will be reassessed when additional data in this patient population become available.
Patients with chronic kidney disease: risk of mortality and cardiovascular morbidity
2 studies5 6 have compared cardiovascular outcomes in patients with chronic kidney disease who were treated with r-HuEPOs to achieve either a high or low haemoglobin concentration. 1 study  showed that patients who were given epoetin alfa to achieve a haemoglobin concentration of 11·3 g/dL had a significantly longer time to the composite endpoint of death, myocardial infarction, hospitalisation for congestive heart failure (excluding renal-replacement therapy), or stroke than did those treated to achieve a haemoglobin concentration of 13·5 g/dL.
The second study compared cardiovascular outcomes in patients with chronic kidney disease and anaemia who were treated with epoetin beta to achieve a haemoglobin concentration of either 10·5–11·5 g/dL or 13–15 g/dL. These groups did not differ significantly in frequency of death from cardiovascular causes, or in time to death from cardiovascular causes or all causes. Trends for all-cause mortality and cardiovascular morbidity consistently favoured the low-target-haemoglobin group, but differences between groups were small. Groups did not differ in the frequency of thrombotic complications.
Both studies[5,6] showed no benefit associated with correction of haemoglobin concentration to high levels compared with low levels. Assuming either the same cardiovascular risk for both treatment strategies, or a slightly lower risk with low target haemoglobin as suggested by the outcome of the first study and given the small trends in the second study there can be little justification for the correction of haemoglobin concentration beyond the minimum level that is compatible with good control of symptoms of anaemia in patients with chronic kidney disease.
A meta-analysis of 9 prospective randomised controlled trials7 assessed all-cause mortality and cardiovascular events associated with r-HuEPO treatment to achieve different ranges of haemoglobin concentration in patients with anaemia due to chronic kidney disease. The results suggest an increased risk of all-cause mortality in patients with anaemia who are treated with r-HuEPOs to achieve a haemoglobin concentration between 12 g/dL and 16 g/dL compared with those treated to achieve a haemoglobin concentration less than 12 g/dL. The best estimate of relative risk of death associated with higher target haemoglobin concentrations was 1·17. No clinical benefit was identified with attaining a haemoglobin concentration higher than 12 g/dL that could not be gained at lower haemoglobin concentrations in patients with chronic kidney disease.
Information and advice for healthcare professionals includes:
- the authorised indication for patients with chronic kidney disease has been changed to stipulate treatment with r-HuEPOs only if symptoms of anaemia are present - all other authorised indications remain unchanged
- dose recommendations have been changed, where necessary, to stipulate a uniform target haemoglobin concentration range of 10–12 g/dL (6·2–7·5 mmol/L) - haemoglobin concentrations higher than 12 g/dL (7·5 mmol/L) should be avoided
- guidance has been provided in the prescribing information for every r-HuEPO for appropriate dose adjustment to maintain haemoglobin concentration within the recommended range
- patients should be monitored closely to ensure that the lowest approved dose of r-HuEPO is used to provide adequate control of the symptoms of anaemia
Prescribing information for r-HuEPOs has been and is being revised to emphasise that treatment to achieve haemoglobin concentrations higher than 12 g/dL (7·5 mmol/L) might increase the risk of death and serious cardiovascular events. Information has also been added about the risk of reduced overall survival and shortened time to tumour progression in patients with cancer who have been treated with r-HuEPOs to achieve haemoglobin concentrations higher than those recommended, and/or to treat patients with anaemia who have not received chemotherapy.
For more information see the public statement from the European Medicines Agency on the conclusions and recommendations of the safety review of epoetins.
See MHRA public assessment reports on erythropoetins r-HuEPOs in patients with cancer and r-HuEPOs in patients with chronic kidney disease.
Article citation: Drug Safety Update December 2007; Vol 1, Issue 5: 2.
Henke M, et al. Lancet 2003; 362: 1255–60 ↩
Leyland-Jones B, et al. J Clin Oncol 2005; 23: 5960–72 ↩
Wright JR, et al. J Clin Oncol 2007; 25: 1027–32 ↩
Bohlius J, et al. Cochrane Database of Systematic Reviews 2006; 3: CD003407 (DOI:10.1002/14651858.CD003407.pub4) ↩
Ajay K, et al. N Engl J Med 2006; 355: 2085–98 ↩
Tilman B, et al. N Engl J Med 2006; 355: 2071–84 ↩
Phrommintikul A, et al. Lancet 2007; 369: 381–88 ↩
Published: 1 December 2007