Study of erythropoietin in treatment of anaemia in patients with rheumatoid arthritis

BMJ 1994; 309 doi: (Published 19 November 1994) Cite this as: BMJ 1994;309:1337
  1. E A Murphy,
  2. A L Bell,
  3. J Wojtulewski,
  4. M Brzeski,
  5. R Madhok,
  6. H A Capell
  1. University Department of Medicine, Glasgow Royal Infirmary, Glasgow G31 2ER
  2. Musgrave Park Hospital, Belfast BT9 7JB
  3. Eastbourne District General Hospital, Eastbourne BN21 2UD
  4. Centre for Rheumatic Diseases, Glasgow Royal Infirmary, Glasgow G4 0SF.
  1. Correspondence to: Dr E A Murphy, Law Hospital, Carluke, Lanarkshire ML8 5ER.
  • Accepted 7 September 1994

Anaemia often occurs in patients with rheumatoid arthritis, and its cause is often multifactorial. The effect of erythropoietin on such anaemia is controversial, and evidence exists that cytokines may affect haemopoiesis, possibly by affecting sensitivity to erythropoietin.1,3 We assessed the therapeutic efficacy of human recombinant erythropoietin in the anaemia of chronic disease in rheumatoid arthritis in a randomised, double blind, placebo controlled study over 20 weeks.

Patients and methods

Twenty patients with definite rheumatoid arthritis and aged 42-75 were enrolled in the study; they were randomly allocated to erythropoietin (10) and placebo (10). Patients had been receiving a stable dose of second line drug treatment for at least 12 weeks. No patients were taking cytotoxic drugs or steroids. Entry into the study required a baseline haemoglobin concentration of <100 g/l (mean of three readings in the month before entry). Iron deficiency was excluded and all patients were given oral iron supplements. One patient with iron deficiency was enrolled in error and was withdrawn. Human recombinant erythropoietin and matching placebo (saline) were supplied in blinded phials. The drug or placebo was injected subcutaneously twice a week for 20 weeks, starting at 40 U/kg (patients 1- 10) and 100 U/kg (patients 11-20). The dosage was reviewed every four weeks and increased if the rise in haemoglobin concentration was <10 g/l per four weeks.


No significant difference existed between the group receiving erythropoietin and the group receiving placebo in terms of age (median 59 (range 42-75) and 61.5 (44-69) respectively). Two patients withdrew from the group receiving erythropoietin. The doses of drug varied from 40 U/kg to 300 U/kg; no patients needed a reduction. One patient developed an occlusion of the retinal vein (in week 11, when his haemoglobin concentration was 115 g/l and packed cell volume 0.352 and he was receiving 40 U/kg erythropoietin).

No significant change occurred in either group in white cell count, platelet count, C reactive protein concentration, rheumatoid factor, blood pressure, visual analogue pain score, Ritchie articular index, duration of morning stiffness, or Stanford health assessment questionnaire score during the study; no differences occurred between the groups for these variables except systolic blood pressure at week 0 (P=0.04, Mann-Whitney U test). A significant fall in the erythrocyte sedimentation rate occurred only in the group receiving erythropoietin (P=0.008, Wilcoxon's matched pairs test). A significant negative correlation existed between haemoglobin concentration and erythrocyte sedimentation rate (r=-0.49, P=0.0031, Spearman's rank coefficient). The Nottingham health profile score for energy improved significantly in the group receiving erythropoietin between weeks 0 and 20 (P=0.028, Wilcoxon's test). Haemoglobin concentration and packed cell volume increased only in patients receiving erythropoietin (figure).


Top: Median haemoglobin concentrations for each group in study. The difference in concentration between the groups was significant at 20 weeks (P=0.02 (95% confidence interval 0.1 to 3.5); Mann-Whitney U test). Bottom: Median packed cell volume values for each group in study. The difference between the groups was significant at 20 weeks (P=0.001 (0.005 to 0.108); Mann-Whitney U test)


This study shows that human recombinant erythropoietin corrects secondary anaemia and results in a clinical improvement in patients with rheumatoid arthritis. In contrast with previous studies of erythropoietin in patients with rheumatoid arthritis,4,5 this study included patients with active disease and receiving modifying treatment - that is, patients in whom persistent anaemia is most difficult to treat. Erythropoietin given twice a week was sufficient for an adequate therapeutic response, giving an increase in haemoglobin by week 4. The fall in erythrocyte sedimentation rate in the group receiving erythropoietin may reflect correction of anaemia as a significant negative correlation existed between the haemoglobin concentration and the erythrocyte sedimentation rate. It is difficult, however, to exclude a modifying effect on the disease of erythropoietin as a trend also existed towards improvement in the group receiving erythropoietin in the values for C reactive protein concentration, visual analogue pain score, and platelet count, although this was not significant. While the expense of human recombinant erythropoietin may limit its clinical use, it may be useful in patients with rheumatoid arthritis undergoing elective surgery or in patients starting second line drug treatment who require faster correction of their anaemia than would be achieved by the modifying effect on the disease of such drugs.

We thank Cilag for providing the human recombinant erythropoietin and matching placebo; metrologists Anne Thomson (Glasgow), Frances McEvoy (Belfast), and Gwynneth Clarke (Eastbourne); Dorothy McKnight for help with computing; and Professor R D Sturrock and Drs A Zoma and Max Field for allowing us to include their patients in the study.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract