The significance of anemia in clinical radiation therapy

Abstract

The time is appropriate for the publication of the excellent review by David G. Hirst of the experimental and clinical evidence for an association between anemia and an increased local relapse rate following radiation therapy.6 The subject needs to be discussed because there is a continuing uncertainty amongst physicians as to whether RBC transfusions, to correct a chronic anemia, indeed decrease the risk of local relapse following a radical course of radiation therapy. Although acute blood loss can be demonstrated experimentally to be detrimental to the oxygenation of tumors, this is not so for chronic anemia. Thus, as Hirst points out, we do not know the mechanism which underlies the clinical association of a chronically low Hb level with an increased risk of relapse after an established protocol for radiation therapy. The systems used in experimental animals do not provide an adequate explanatory model for the complexity of the clinical situation. Hi&s review provides references to the clinical reports which document the prognostic significance of Hb level for the control of malignancies in patients undergoing radiation therapy. The patient groups in these reports are restricted to those with cancer of the cervix or of the head and neck. Note, criteria for defining a low Hb varies with the report. Some Hb levels at which an increased rate of local relapse has been noted lie in what otherwise might have been classed as the normal range. Nevertheless, there is a heavy weight of clinical evidence, particularly for those patients with cancer of the cervix, which indicates that, given a patient population with a range of Hb levels during treatment, the largest proportion of patients with local relapses will be in those with a Hb level in the low end of that range. 1 will illustrate the point with data from the Princess Margaret Hospital (PMH). Together with my associates at the PMH I have previously reported that for patients with locally advanced cancer of the cervix the average Hb level during radiation therapy is a significant prognostic factor.2*3 How powerful a prognostic factor is that Hb level? To show the relative power I have ranked the Hb level using a Cox regression analysis.4 Used in the analysis are all those patients with Stage IIB or III cancer of the cervix seen and treated with radiation therapy between the years 1965 and 1975 at the PMH. Patients who did not receive both intracavitary 13’Cs and external radiation therapy have been excluded. By keeping the analysis to the years 19651975 all patients will have been followed 10 years or longer. Also, the treatment protocols varied little over that period. The prognostic factors ranked in the Cox regression analysis were Stage (IIB v. III), age (~50 v. r50), central dose (~7000 v. 27000 mixed rads), lateral pelvic dose (~4500 v. 24500 mixed rads), and average Hb level during treatment (~10 v. 10-l 1.9 v. 12-13.9 v. rl4gm%). For both local relapse and overall relapse, stage is ranked first and Hb level second with both having a p value of less than 0.000 1. Age was just significant at a p value of 0.046. With the exclusion of patients with incomplete treatment as noted above, central and peripheral doses were not significant in this particular analysis. Clearly, then, the Hb level is a significant and independent prognostic factor for patients with cancer of the cervix, Stages IIB and III. The question to be answered is whether a low Hb level is just a marker of disease which has an inherently poor prognosis or whether it predicts a poor prognosis because of the relative inefficacy of radiation therapy when a low Hb is present during treatment. To provide information which might help to answer the question just posed, I carried out a log-rank analysis of the same patient data described above, adjusting for stage, central dose, side wall dose and age, and determined the relative control rates at different Hb levels.* Shown in Table 1 are the relative rates for local relapse (LRR), distant relapse (DRR), and total relapse (RR). As can be seen for these patients with Stage IIB and III cancer of the cervix, the major influence of the Hb level during treatment is on the LRR and RR and not the DRR. The LRR decreases from a high of 46% for Hb levels less than lOgm% to 20% for those with a Hb level greater than 12gm%. Note that the effect of bulk of disease as defined