Clinical treatment for anemic patients on hemodialysis (HD) has significantly improved since the introduction of erythropoiesis-stimulating agents, such as epoetin alfa (EPO). Treatment with EPO is effective in raising hemoglobin (Hb) levels in anemic patients to the target range (10-12 g/dL) determined by the United States Food and Drug Administration. However, maintaining a target Hb level during EPO treatment is difficult. Some studies have indicated that only 10% of HD patients receiving EPO treatment are able to maintain stable target levels from month to month. Recently, a national, retrospective study examined the relationship between Hb variability and mortality.

The study assessed 159,720 HD patients who had EPO claims in each of the first six months of 2004. This time frame was considered the study’s exposure assessment period during which Hb levels were monitored. After these six months, patients were followed until the end of 2004, or until death or change in dialysis treatment. The study utilized two classification systems to define Hb variability; one based on monthly Hb values, and the other based on the lowest and highest Hb values observed during the exposure assessment period. Hemoglobin categories within each system were defined as follows: low (Hb < 11 g/dL), intermediate (Hb = 11-12.5 g/dL), and high (Hb > 12.5 g/dL). Descriptive statistics were used to assess the classification system’s various categories, with adjustments made for demographic and comorbidity characteristics.

Patients with Hb values in the intermediate and high ranges showed no increased risk for death, while patients consistently in the low category had the highest risk of death. Also, patients with low Hb levels in the first 3 months of the assessment period and higher Hb levels (> 11 g/dL) in the second 3 months had lower mortality risk than patients with low levels in the second 3 months. In general, more patients with low Hb levels in the second 3 months indicated a higher risk of mortality.

These results suggest the number and timing of low Hb levels are the most important factor in terms of mortality risk. Patients who experienced falling levels during the latter months of the assessment period were at a significantly higher risk for death than patients whose Hb levels were increasing in the latter months. In general, the model indicated that > 3 months with Hb levels < 11 g/dL may be associated with increased risk of death. The results also address concerns about mortality risk with Hb levels above the target range. Gilberson et al concluded this study did not find increased mortality associations with higher Hb levels in patients, and indicates the need for future trials to examine optimal EPO dosing for HD patients.

Hemoglobin Level Variability: Associations with Mortality. Gilbertson DT, Ebben JP, Foley RN, Weinhandl ED, Bradbury B, Collins AJ. Gynecol Oncol. 2007 Nov 23.

NAAC Expert Commentary:
Anemia management is a key component of chronic kidney disease care and one that has undergone substantial scrutiny. Whereas target hemoglobin (Hb) ranges have been recommended for patients receiving erythropoiesis-stimulating agents (ESAs), there is prior evidence that patients regularly experience fluctuations outside the range, so-called hemoglobin cycling. In this retrospective study, the authors sought to better define the impact of fluctuating Hb values. The findings which ultimately proved most interesting were the very frequent nature of Hb cycling and the association between Hb values below the K/DOQI target and the risk of mortality in patients receiving hemodialysis.

The implications of the data are that recognition of Hb cycling should be incorporated into the design of quality improvement projects tied to anemia management, and that clinicians should follow current treatment recommendations regarding the lower limit Hb target value. The absence of an association between either Hb cycling or a high Hb value and patient mortality, despite previous publications which have suggested otherwise, highlights the need for additional studies on these issues and others (e.g. impact of ESA dosage vs. Hb achieved on patient morbidity and mortality), in the dialysis population.

As with any observational study, limitations must be considered when applying relevance to the results; in this case those limitations which could help explain the Hb variability. Limitations recognized by the authors included the likely presence of confounding-by-indication biases, the exclusion of patient data from all those who did not have a complete dataset, and the lack of complete comorbidity profiles addressing issues such as hospitalization and infection. These factors will undoubtedly be addressed in future prospective clinical trials intended to help optimize the treatment of anemia.

Last Modified: March 6, 2008