London, UK - Risk of all-cause mortality was significantly higher in patients with anemia caused by chronic kidney disease (CKD) who were treated with erythropoiesis-stimulating agents (ESAs) to higher rather than lower hemoglobin targets, according to a meta-analysis of nine randomized controlled trials and more than 5000 patients [1].

The study, which was published in the February 3, 2007 issue of the Lancet, also found that patients in the higher-hemoglobin-target group were at increased risk for arteriovenous access thrombosis and poorly controlled hypertension.

Does this put an end to the debate about target hemoglobin levels in CKD patients with anemia? It does provide evidence about upper levels of hemoglobin and cardiovascular risk, but other questions — such as use of ESA dose vs hemoglobin-target levels—remain to be answered.

"The most important of our findings is the significant increase in the risk of all-cause mortality seen when a target hemoglobin concentration of 120 to 160 g/L is achieved," despite this being within the normal physiologic range, the researchers, led by Dr Arintaya Phrommintikul (Monash University, Melbourne, Australia), write.

The group explains that lower hemoglobin levels in patients with CKD have been linked to decreased energy, cognition, exercise capacity, and quality of life and increased mortality risk. They add that although treatment with ESAs, such as epoetin alfa and beta and darbepoetin alfa, has been linked to improved exercise capacity and cognition in some small studies, the ideal hemoglobin treatment-target level remains very controversial. The November 2006 publication of two major trials—the Cardiovascular Risk Reduction by Early Anemia Treatment with Epoetin Beta (CREATE) [2] and the Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) [3]—further raised the issue of cardiovascular risk.

Therefore, the researchers performed a meta-analysis to determine whether ESA treatment targeting different hemoglobin concentrations in these patients is linked to changes in all-cause mortality and cardiovascular events.

They searched the English literature to November 2006 for prospective randomized controlled trials of more than 100 patients lasting longer than 12 weeks that assessed hemoglobin targets in this population. Nine trials, with 5143 patients, met their criteria. Patients received epoetin alfa in eight trials and epoetin beta in one trial (CREATE).

They found significantly higher risks for all-cause mortality, arteriovenous access thrombosis, and poorly controlled blood pressure in the higher-hemoglobin- than in the lower-hemoglobin-target group.

This is the first demonstration of significantly increased all-cause mortality with higher hemoglobin targets achieved with ESAs in patients with anemia caused by CKD, corresponding author Dr Henry Krum (Monash University) said.

The group writes that the mechanisms that underlie the increased mortality remain unclear. The Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT), which is investigating darbepoetin, will provide further insight into optimal target levels, they note.

Krum listed some of the many unanswered questions: By what mechanism does raising hemoglobin with ESA increase events? What is the optimal hemoglobin level in CKD patients? Do differing ESA formulations matter? Do these results apply to other conditions (eg, chronic heart failure) in which hemoglobin is low?

The team concludes that this meta-analysis raises questions about the appropriateness of current target hemoglobin concentrations; they suggest that future guidelines should consider an upper limit for target hemoglobin concentrations.

In an accompanying editorial [4], Dr Giovanni FM Strippoli (Consorzio Mario Negri Sud, Santa Maria Imbaro, Italy, and part of the Cochrane Renal Group, University of Sydney, Australia) and colleagues write that the findings by Phrommintikul et al—the "dramatic" increase in the risk of several end points with higher vs lower hemoglobin targets—is consistent with results from earlier trials. Collectively, these show a net clinical harm of targeting higher hemoglobin concentrations, which was identified nearly 10 years ago by Dr Anatole Besarab (Henry Ford Hospital, Detroit, MI) and colleagues [5], the authors write, adding that major methodological pitfalls weaken the claim of other trials that patients show quality-of-life benefits with complete normalization of hemoglobin levels.

The editorialists conclude: "The question has been answered: higher hemoglobin-target concentrations increase mortality via cardiovascular end points. Part rather than complete correction of anemia is appropriate, although commercially less attractive, and it is time to move on."

Strippoli noted the clinical implications: "Treat patients with CKD to a hemoglobin target of 100 to 120 g/L; if patients have severe cardiopathy, treat them to a hemoglobin target of 100 to 105 g/L." He also discussed the implications for research: "First, hemoglobin-target trials should be stopped, or their aims need to be strongly reassessed; second, a trial of different doses of erythropoietin is needed, because what is causing harm to the patients is probably that many of them do not respond to these drugs properly and higher doses are used."

Dr Jeffrey S Berns (University of Pennsylvania School of Medicine, Philadelphia), who was not involved in this study, commented that in aggregate, these studies argue that hemoglobin levels in patients with CKD should not be maintained above the range of 12.5 to 13 g/dL. Pushing down on the upper limit of a target hemoglobin range is almost certainly going to increase the number of patients with hemoglobin levels below 10 to 11 g/dL, he added, noting that this makes results from the ongoing TREAT study important. He summarized: "We seem closer to defining the upper limits of what the hemoglobin level should be; the lower limits still need to be explored. The risks of being closer to 13 g/dL compared with below 10 g/dL also need to be clarified."

Dr Adeera Levin (University of British Columbia, Vancouver) said that the meta-analysis and editorial comments in the Lancet "add to the current flurry of papers describing the issue of target hemoglobin values in patients with kidney disease."

She noted that it was a well-done study of available trials, but that it would have been interesting to determine whether the same harm could be demonstrated in the nondialysis populations alone; also, some of the CKD trials had methodological flaws (eg, differential comorbidities at baseline in the CHOIR study).

When asked whether it is ready to be accepted that higher hemoglobin values are more likely to cause harm than benefit and to cease all trials examining this question, she notes: "As a nephrology community, we continue to be struck by the heterogeneity of the populations we care for, and we should seriously question the notion of 'one size fits all,' " adding that perhaps the studies to date have been too naive and the guidelines too simplistic. She concludes: "The consensus within the community, even with this meta-analysis, is not that no trials are needed, but rather that appropriately designed trials, examining more fundamental issues, as suggested by Strippoli and colleagues (fixed-dose ESA vs targeted hemoglobin), to determine whether there is a benefit of replacing a hormone or reaching a target, should be undertaken. The debate is not yet over, but perhaps the simplistic approach to asking the questions may be. "