Additional Concerning News About EPO Treatment

Abstract & Commentary

Synopsis: In the current report, a disconcerting finding of a strikingly increased incidence of thrombosis in women with uterine cervical or vaginal carcinomas in whom erythropoietin was used to maintain hemoglobin levels at 12 g/dL was presented. The report was a retrospective review of nonrandomized patients treated at a single institution and accordingly runs the risk of over-interpretation. Nonetheless, the findings are quite dramatic and clearly warrant careful review and analysis by well-constructed clinical trial methodology.

Source: Wun T, et al. Cancer. 2003;98:1514-1520.

Wun and colleagues from the University of California-Davis performed a retrospective, case-control study on consecutive patients with localized carcinoma of the uterine cervix or vagina who were treated with chemotherapy and radiation, examining for factors associated with the development of symptomatic venous thrombosis. Based upon the finding that patients with this malignancy who had their hemoglobin levels maintained by transfusion had improved clinical outcomes,1 it was the practice at this institution to either transfuse or treat with recombinant human erythropoietin at the earliest signs of anemia.

Records from 147 patients were reviewed. These represent a consecutive series of patients treated with chemo-radiotherapy for localized cervical or vaginal carcinoma (FIGO Stage IB-IVA) at that institution from 1994 to 2002. Only patients who were treated with combined, concurrent chemotherapy and radiation were included, and patients who underwent surgery as part of their therapy were not excluded from the analysis. Radiation included both external beam and intracavitary doses, and chemotherapy included cisplatin and 5-fluorouracil by continuous infusion for 4 days during weeks 1 and 5 of radiation therapy (in the majority of cases).

During the early phase of the analysis period (1994-1999), hemoglobin levels were maintained at > 11 g/dL, primarily by red cell transfusion. After 1999, recombinant human erythropoietin (rHuEPO) was routinely used whenever hemoglobin levels fell below 12 g/dL. Patients were considered to have had a symptomatic thrombotic event if 1) they presented with signs and symptoms consistent with either upper or lower extremity thrombosis; and 2) there was objective, radiographic verification of the thrombosis, usually by Doppler ultrasound.

Analysis revealed that there were no significant differences in age, disease stage, or body mass index between those women who received rHuEPO (n = 75) and those who did not (n = 72). There were 17 episodes of symptomatic thrombosis in the rHuEPO-treated patients compared to only 2 episodes among the 72 patients who did not receive rHuEPO. Thus, patients who received rHuEPO had an odds ratio (OR) of developing thrombosis of 10.3 (95% confidence interval [CI], 2.3-46.2). Multiple logistic regression revealed that only the use of rHuEPO was associated with an increased risk of thrombosis (OR, 15.3; 95% CI, 3.1-76.7).

Comment by William B. Ershler, MD

The news here, of course, is of great concern. However, before we make wholesale changes in the way we treat anemia in chemotherapy- and radiation- treated cancer patients, we need to take a critical look at the data presented.

First of all and most importantly, this was a retrospective, nonrandomized review from a single institution. The patient groups (those that did vs those that did not receive rHuEPO) may well reflect 2 distinct populations that were different in parameters not immediately obvious. Wun et al carefully examined those demographic features known to be associated with increased clotting (such as body mass index and age), but others certainly may exist. Patients in the rHuEPO group, many of whom (23 of 75) also received RBC transfusions, may have had more extensive or locally aggressive disease than those in the no-rHuEPO group, of whom 31 of 72 did not receive either rHuEPO or transfusion. Thus, just over half of the patients in the non-rHuEPO group had anemia, compared to 100% of those who received rHuEPO treatment. The presence of anemia associated with treatment might well indicate more advanced disease or reduced constitutional factors predisposing to thrombosis.

The occurrence of thrombosis in erythropoietin-treated patients is not a new observation. In patients with renal failure and anemia, rapid correction with rHuEPO is a known risk factor for thrombosis, and, accordingly, the package inserts for epoetin and darbepoetin both include warnings to this effect. Nonetheless, thrombotic events had been considered less common in rHuEPO-treated cancer patients, although in general, this population is known to have an increased incidence of clots. Thus, the current report is a bit of a wake-up call for medical oncologists to be aware of this untoward effect and to treat with caution. Short of findings indicating that patients with gynecological malignancies are particularly susceptible to this adverse effect of erythropoietin, which is unlikely, clinicians should extrapolate this caution to all patients treated with this growth factor. Conventional wisdom would suggest that we "start low and treat slow," not unlike what nephrologists have been doing for years. Further investigation may indicate that those at greatest risk are the patients for whom anemia is corrected rapidly.

Dr. Ershler, INOVA Fairfax Hospital Cancer Center, Fairfax, VA; Director, Institute for Advanced Studies in Aging, Washington, DC.


1. Grogan M, et al. Cancer. 1999;86:1528-1536.