L. Tim Goodnough, MD: Thank you. As was mentioned, I am director of the blood bank at Stanford, but I come to you as a hematologist/oncologist and I have had 25 years of experience with patients with solid tumors and leukemias, stem cell transplantation, including the intensive care unit, and operating room privileges going in and doing bone marrow harvests. So I have hung my share of blood, and I would like to think I know what goes on at the bedside.

Paradoxically, even though I sell blood for a living, my academic interest has been in devising strategies for patients to avoid a blood transfusion, and so the topic that I would like to share with you tonight, sort of at the 30,000-foot view, is the prevalence of anemia in certain clinical conditions, and the relationship between anemia and clinical outcomes. We will be focusing on chronic kidney disease and cancer tonight.

Now, I just want to start first with a so-called elderly population. I used to think 65 and older was elderly, but it is not so old to me now, and as shown here, the prevalence of anemia is quite high, approaching 11%, for all patients over the age of 65, and even more so in minority populations such as non-Hispanic blacks and Mexican-Americans.

Now, if you do a workup of these patients who are anemic, it is informative to understand that a third is going to have nutritional deficiencies, and so this is, I think, an important take-home message, that the first thing we should be thinking about is iron deficiency in the elderly who are anemic, either because of gastrointestinal (GI) blood loss or nutritional causes, but also B₁₂ and folate or combinations of these have to be considered.

The second largest category of anemia, besides nutritional, would be chronic diseases, either chronic kidney disease, or the anemia of chronic inflammation, together comprising about 30%, again about a third of the patients, and then the remaining third are classified as what we would call undifferentiated. My bias is that a lot of these may in fact be chronic kidney disease because it is underappreciated that a normal creatinine level in an elderly individual with a low body mass may in fact represent pretty significant renal dysfunction, and you have to go through glomerular filtration rate calculations to be sure, and also primary hematologic diseases such as myelodysplastic syndrome. So, I think that anemia is greatly underappreciated in the population as a whole, but certainly in the elderly in particular.

It has significant clinical impacts, as you might imagine, in the ability of the elderly or the extreme elderly, over the age of 75, in terms of maintaining their mobility and physical performance, and actually their survival, so it does not take much for an ambulatory extreme-elderly patient to become nonambulatory because of reversible causes such as anemia, and this should receive careful consideration.

Evolving questions in this setting are substantial and the National Institutes of Health (NIH) in particular has a series of funding initiatives to address these issues, such as should there be a different biologic definition of anemia in the elderly compared to the nonelderly, and what would be the pathophysiology of the unexplained anemia in the elderly? Would it be because of decreased marrow cellularity; impaired erythropoietin response; ineffective erythropoiesis, such as myelodysplastic syndromes; or inflammatory mediators? Is aging itself an inflammatory state, and most importantly, can anemia management improve clinical outcomes in this group?

I next want to turn to the pathophysiology of the anemia of chronic disease, and this has been a very fertile area of advancement in the last several years. As you know, it is a multifactorial process, and certainly there is diminished erythropoietin production in the anemia of chronic inflammation.

There is direct inhibition of erythroid progenitor cell production, and most tellingly, impairment in iron metabolism, and it is this third symptom where there has been the most advancement in our knowledge of the anemia of chronic disease.

This is a complicated table from a New England Journal of Medicine review that we published two years ago with the title, "Anemia of Chronic Disease" emphasizing how difficult it is for clinicians to sort out patients who have chronic disease alone, or in combination with iron deficiency. So if you take a look at traditional biochemical parameters, such as iron, transferrin, transferrin saturation, and ferritin, in the anemia of chronic disease, and compare them to patients who have chronic disease and iron deficiency, it is very difficult to sort these 2 categories out, and as a hematologist, I would not be able to tell you the iron status of your patient with chronic disease, short of doing a bone marrow and looking for stainable iron. And there have been various schemes to try to get around this by looking at an assay called soluble transferrin receptor and taking the ratio of that to the log of ferritin. As you would imagine, this is not very practicalfor the practicing physician—getting a laboratory report that does not go through these calculations for them—but this has been postulated as one way to separate out anemia of chronic disease versus chronic disease in the presence of iron deficiency.

The story that is most interesting to me is the hepcidin story, and this was identified by Elizabeth Nemeth and John Gantz, from the University of California at Los Angeles not so many years ago, where they demonstrated in a rodent model that if you give a rodent an infusion of interleukin-6 (IL-6), which is an inflammatory cytokine, over 3 hours, shown here with peak levels, very shortly after that, hepcidin, which is an inflammatory peptide, is produced very quickly, measurably, in the urine, within 2 hours of this infusion. And it is accompanied by an acute reduction in serum iron, hypoferremia, and reduced iron saturation, and teleologically, having anemia of chronic disease in inflammation has been explained as a defense mechanism to deny iron to pathogens.

This is a picture from our review in the New England Journal of Medicine showing how this works in the setting of a pathogen, which will set off a cascade of inflammatory cytokines, such as gamma interferon, tumor necrosis factor, IL-1, and IL-6, which I showed you in the previous slide. IL-6, for example, is stimulating hepcidin. It was demonstrated to be produced by the liver, which is why it is called hepcidin, and along with tumor necrosis factor, inhibiting erythropoietin production, and also inhibiting red blood cell precursors, leads to a direct inflammatory suppression of red blood cell production.

What hepcidin does is act at 2 sites. It gives you an immediate reduction in iron absorption and an immediate sequestration of iron in the macrophages, which is why we have this acute hypoferremia, and this presumably denies iron from pathogens as a defense mechanism.

We were asked by the New England Journal of Medicine to come up with an algorithm for the lay physician on how to approach anemia in their patients, and this is what we tried to illustrate. If you have anemia in the setting of inflammation, either by diagnosis or with some kind of a biochemical marker, and you have a low transferrin saturation and you have ruled out other causes of anemia, next go to your ferritin. In the setting of a patient, if the ferritin is less than 30 ng/mL, almost certainly your patient is iron deficient. If the patient has a ferritin of more than 100 ng/mL—and I have been criticized that perhaps this is too low, and some people advocate 200 ng/mL—if it is more than 100 to 200 ng/mL, you are probably pretty sure that you are not dealing with iron deficiency, although you can not absolutely rule that out because ferritin is an acute-phase reactant, and in the setting of inflammation, you lose it as a road map as a marker forstorage iron, but if your ferritin is more than 100 to 200 ng/mL, think anemia of chronic disease. Between 300 and 100 to 200 ng/mL, think of a combination of chronic disease and iron problems, and you can either go through a transferrin receptor assay and divide it by the log ferritin, or more probably, in my view, an empiric trial of iron, not forgetting to rule out colon cancer and other things that need to be entertained, but in terms of anemia management probably the most intelligent way to approach this would be an empiric trial of iron to see if the anemia is improved.

Now, we also were asked to give a table on the therapeutic options for treating the anemia of chronic disease, and it is important to remember that you need to treat the underlying disease. This was illustrated in rheumatoid arthritis, which is your classic prototype for inflammatory disease, and antitumor necrosis factor, or infliximab, has been available, and when you give that to patients with rheumatoid arthritis, their anemia corrects, and when you address the inflammation, that is part of anemia management. Transfusions are indicated if the patient is unstable or has symptoms or is severely anemic, for example, less than 6 g/dL, and that would be appropriate both in the anemia of chronic disease, and the anemia of chronic disease with true iron deficiency. I almost cringe to say this—to me a transfusion is contraindicated when a patient has iron deficiency—but if they are unstable, transfuse.

Iron supplementation should be given if you are convinced they are iron deficient. It should not be given in the setting of chronic disease alone although this, I think, deserves further study. The question is, should we be treating anemia of chronic disease and iron sequestration syndrome, with a combination of iron and erythropoietic agents, and I think that that is an area that has been investigated in oncology and needs to be investigated in other inflammatory states, such as chronic heart failure and in the critical care setting, which you will hear about later.

Erythropoietic agents are indicated, as you all know, for certain diagnoses in chronic disease, and you should also use these in chronic disease after you have ruled out a response to iron.

I have tried to make the case over the last several years that anemia should not be an abnormal lab value. My view on this is that it is an unmet medical need with serious clinical consequences, and needs to be paid attention to and needs to be corrected. What I would like to do is go through the data to try to convince you of that.

The adverse outcomes in patients with anemia are, first and foremost, blood transfusions. I have not yet met a patient who is excited about getting a blood transfusion. This is a very patient-friendly thing to do, to devise strategies to help them avoid blood transfusions, and interestingly enough, in the whole field of blood conservation, as you are aware, the Food and Drug Administration (FDA) has always bought into the fact that if you can show a blood transfusion-sparing effect, that is an indication for a label for that agent in a particular clinical setting.

But increasingly we are learning, and I will go through some of this data with you, and there are adverse effects on cardiac function, and morbidity and mortality in patients with impaired cardiac function. Quality of life has been well documented to be improved with anemia management in patients with chronic kidney disease and patients with cancer, and for a while we were hoping to be able to demonstrate that with anemia management you could fundamentally affect disease progression, and improve patient survival. As you are all aware, in the last several months, there have been some clinical trials that have demonstrated that is probably not the case and I will be addressing that towards the end of the session.

Let's first turn our attention to the anemia of chronic disease.

These are data from Medicare—this is a Medicare registry study, published in the late 1990s, but it is actually data from the early 1990s—and what they looked at was all-cause deaths, deaths from cardiac, and deaths from infectious disease, the 2 most common causes for death in the dialysis population, and they looked at 1992 and 1993, and stratified the odds adjusted ratio of death, according to the hematocrit cohort that year. So they set at one the likelihood of death for the hematocrit cohort from 30% to 33%. And as you can see, with increasing degrees of anemia, there was a 12% and a 33% increase in mortality. Now obviously this is a retrospective observational study, but it was an enormous Medicare database, and it was highly statistically significantly different. They were not able to show a reduction in mortality that was statistically significantly different at higher levels of hemoglobin because of the relatively small number, but when they combined1992 and 1993, they were able to show a reduction in mortality that was statistically significant and so it was this that led the National Kidney Foundation (NKF) to develop guidelines, first in 1994, updated in 1997, and again last year, to actually lead their field in anemia management.

Why did they do this? It is because it was an underrecognized problem. Eight percent of the US population has abnormal serum creatinines; from some studies we know this. At the time, only 7% of chronic kidney disease was treated with erythropoietin therapy and we now know that long-term anemia is a risk factor for surrogate markers, such as left ventricular hypertrophy, which itself is a risk factor for mortality.

The anemia affects approximately 20 million adults in the United States, and a large US multicenter study of predialysis chronic kidney disease is the kind of patients you are going to be seeing, or you do see, of 5000 patients found that anemia was present in half of these cases.

The NKF's Disease Outcomes Quality Initiative (DOQI) guidelines, which is sort of a manual from A to Z on how to take care of dialysis patients and patients with chronic kidney disease, were upgraded this last year for both predialysis and dialysis patients. They said hemoglobin should be maintained between 11 g/dL to 12 g/dL in this setting, and I think this is a wonderful example of who to lead a field instead of simply sitting back and reflecting what the practice is. When they first started out in 1994, very few nephrologists were maintaining their dialysis patients between 11 g/dL to 12 g/dL, and subsequently they have demonstrated that this is a benefit in the predialysis population also.

What is noteworthy is that you can get to this level with a variety of dosing strategies, so this is the percentage of patients on the Y axis who successfully maintained hemoglobins greater than 11 g/dL by the dosing strategy on the X axis. So the first bar shows that when you take all dosing strategies together, 80% of patients with chronic kidney disease can be maintained at hemoglobins of greater than 11 g/dL. Probably the least effective way to do this is once-weekly dosing strategies, but as you can see, these bar graphs range between 70% and 90% and the message is that you can treat patients weekly, you can treat them every 2 or 3, or every 4 weeks, or even less often than that, and with the proper dose, with these intervals, you can get to where you want to go.

Here is one such study, for example, from Spinowitz, that was just presented as a poster last month. It examined at the value of an extended dosing regimen for the initiation of epoetin alfa for the treatment of anemia of chronic kidney disease, and again, on the Y axis they looked at the percentage of patients who could be maintained at hemoglobins of greater than 11 g/dL, or an increase of greater than 1 g/dL from baseline. As you can see— the yellow bar is representing hemoglobins greater than 11 g/dL—they were able to successfully extend the dosing interval all the way out to every 4 weeks and maintain the efficacy of getting to a hemoglobin of 11 g/dL.

The other thing that you should know is that even when you transplant patients with chronic kidney disease, there is a certain percentage of patients who continue to have problems with anemia posttransplant, so this is a retrospective cohort study of about 240 patients. The proportion of patients, for example, with hemoglobins of less than 9 were 76% at transplantation, went down to 21% 1 year after transplant, and back up to 36% 4 years posttransplant, for a variety of reasons: chronic inflammation, myelosuppressive or immunosuppressive therapy and so forth. I think what we need in this population is continued vigilance, better anemia screening, and long-term management, even after transplant.

What about anemia and congestive heart failure? We know that anemia is associated with worse symptoms, greater impairment in functional capacity, and a significant increase in mortality in patients with advanced heart failure. When you manage the anemia by increasing the hemoglobin as little as 1 g, this is associated with a reduction in length of stay, decreased hospital charges, and a reduction in mortality. Every 1 g decrease in admission hemoglobin values, in patients with significant heart failure, is associated with a corresponding increase in death or rehospitalization, so I think that there is a growing awareness that anemia is again not an abnormal lab value, but a serious comorbidity and needs to be incorporated into heart failure treatment guidelines, which to my knowledge are not yet part of the heart failure guidelines. If you take a look at the guidelines, getting a hemoglobin and paying attention to the hemoglobin is simply not on the radar, and that needs to beaddressed.

Here is an example of how interesting this is from a hematologic point of view. This is the survival over 1 year of class III/class IV New York Heart Association heart failure patients by quartile of hemoglobin. The pink is the upper most quartile, the yellow is the lowest most quartile, and as a hematologist, I would not even regard the upper 3 quartiles as anemic. But, with reduction in hemoglobin by quartile, there is a substantial reduction in survival from 80%, with hemoglobins of more than 14.8 g/dL to 60% in hemoglobins of less than 12 g/dL.

Is this true and unrelated or is it cause and effect? We do not know from studies like this, but we do know that when you look at the impact of hemoglobin level and age on length of stay in patients with heart failure there is a direct correlation, so these are bar graphs of patients aged less than 70 years, 70-79 years, and greater than 80 years, and all ages here, and this is mean adjusted length of stay. As you can see, for those who are least anemic or not anemic, this would be the length of stay and there is a corresponding increase with increasing degrees of anemia.

And it has a corresponding effect on patient charges shown here.

Anemia has also been shown to be a predictor of death or rehospitalization in patients who do have decompensated heart failure, so this would be death or rehospitalization, the rates with normal hemoglobins or increasing degrees of anemia.

Clinical outcomes by hemoglobin in patients with decompensated heart failure are shown here by hemoglobin level, normal, and increasing degrees of anemia, and this is inpatient, in-hospital patient mortality, also for 60-day mortality and for rehospitalization.

We have looked at this and the question is, is this an association or cause and effect? So this study attempted to address this by randomizing patients to no anemia management versus an anemia-management strategy.

This is a relatively small study from Israel. This is the control group. Look at the hemoglobin levels in these patients with class III/class IV heart failure. The severely anemic, compared to the Horwich study I showed you by quartile. So when you do nothing over 9 months, in terms of anemia management, the hemoglobin does not change. The left ventricular ejection fraction, if anything, goes down. The need for Lasix® (furosemide) therapy goes up, and the number of hospital days goes up. When you treat these patients with a combination of intravenous iron and epoetin alfa your hemoglobin goes from 10 g/dL to 13 g/dL, or left ventricular ejection fraction, if anything, goes up, the need for Lasix® (furosemide) therapy goes down, and the number of days in hospital goes down. There were mortalities in this group over 9 months. There were zero mortalities in this group. It was a relatively small study—I can not say too much about the statisticalassociation—but in terms of proof of principle, when you manage anemia I think clearly there is an indication that you reduce morbidity and perhaps mortality, and this is being followed up with randomized controlled clinical trials to demonstrate that point.

Anemia in myocardial infarction.

There was a very large Medicare registry study by Wu published in the New England Journal of Medicine. The next slide is very complicated, but I want to walk you through this.

When you do a retrospective observational study and you compare transfused patients to nontransfused patients, transfused patients always do worse, no matter how carefully you control for the known comorbidities as a physician, when you make a decision to transfuse that patient, but not that one, my view on this is the transfused patient was inherently sicker.

If you take a look at all patients and you set the odds adjusted ratio of in-hospital mortality at 1, and these are patients over 65, in the coronary care unit, with a primary diagnosis of MI, nobody went to surgery, those were excluded, so we are excluding surgical blood loss, and you look at the transfused versus the nontransfused cohorts, there was a 48% increase in mortality in the patients transfused. And if you now look at it by admission hematocrit—these are the patients who came in not anemic, again, the transfused patients did worse than the nontransfused. With increasing degrees of anemia, these mortalities are still higher in the transfused versus the nontransfused, but it is starting to telescope together here, and with admission hematocrits that are significantly anemic at admission, with a transfusion, you have a substantial reduction in mortality compared to those not transfused. Now, there are a lot of problems with this study and Lena would be happy tohelp you understand that in the question/answer period, but for example, there is a lot we do not know. We do not know what the hemoglobins were at the time of the transfusion. We do not know why they were transfused, how much blood they transfused, and there is apparently an imbalance in the percentage of patients who are in Inderal® (propanolol) therapy in the one cohort versus the other, but I think it is instructive that this is the only observational retrospective trial I am aware of where it shows that a transfusion is actually associated with a better outcome than not getting transfusion. So I think we need to be cautious about patients with acute coronary syndromes in the coronary care unit, with transfusion trigger philosophies, and I think that the other speakers are going to be addressing this in their talks.

What can we say about the relationship between anemia and cancer and cancer chemotherapy?

If you take a look at radiation therapy patients, for example, by category of disease, many of them are anemic at baseline, before radiation therapy, and then during radiation therapy the prevalence of anemia goes up. So it is a common clinical problem.

Previous studies had demonstrated that in patients who can be demonstrated to have oxygenated tumors versus hypoxic tumors, there is a much better disease-free survival at 1 year. That goes along with how we understand the clinical benefit of radiation therapy requiring oxygen radicals for benefit.

They were able to tie this together in patients with cervical cancer by level of hemoglobin at the time of radiation therapy and the likelihood of recurrence. So if they had hemoglobins of more than 12 g/dL during treatment, their local recurrence rate was much lower than if they were anemic, as was survival.

Here is another clinical setting, in squamous cell carcinoma of the head and neck, where the 2- and 5-year survival was very much affected by whether the patient had normal hemoglobin levels, higher survival rates, than the patients who were anemic.

And correspondingly, in the chemotherapy literature, Littlewood demonstrated in a very large study of a variety of solid and liquid tumors, that if you randomized patients to placebo versus epoetin alfa and raised hemoglobin levels, the probability of survival over time was actually better in the erythropoietic agent group than in the placebo group. It is these 2 clinical observations in radiation therapy and chemotherapy that led to a series of clinical trials in an attempt to normalize hemoglobin and have an impact on overall survival.

There is a recent Cochrane analysis on a large number of randomized clinical trials, a systematic review and an economic evaluation of the erythropoietic agents in cancer to see whether there was efficacy and safety. What they demonstrated was that yes, the erythropoietic agents, in cancer-induced anemia, reduce blood transfusion by an estimated 18%, and if you remember, this is why we received the labeled indication for these agents as a blood transfusion-sparing event. So they concluded that erythropoietic agents are effective in improving hematologic response and red blood cell transfusion requirements, and appear to have a positive effect on health-related quality of life.

As you are all aware, over the last 3 to 4 months, there has been a flurry of interest in these series of clinical trials I alluded to. There are 6 of them in a variety of settings—including breast cancer, non–small-cell lung cancer, undifferentiated cancer not receiving chemotherapy, and chronic kidney disease—in an attempt to show, with normalization of hemoglobin, that you can improve survival. To make a long story short, they were unsuccessful at showing that, and in fact, the data seems to indicate the opposite. So as of March 9, there have been new black box warnings in cancer and chronic kidney disease, and other speakers will address perioperative surgery.

To make a long story short, if you remember nothing else from tonight's symposium from my summary, the hemoglobin levels should be less than or equal to 12 g/dL and no higher, and the FDA is very concerned that we not go above that level, and reminds everybody that the reason why these agents should be used is a blood transfusion sparing effect. Overcorrection of anemia above these levels is associated with adverse outcomes.

I am not going to go through the 6 studies, but I just want to just briefly mention 2 of the studies just to give you a flavor for where we are coming from. You have probably read about these studies in the New York Times and the Wall Street Journal, which is where I get all of my news. One of them was in head and neck cancer, it was an open-labeled randomized trial, comparing radiation therapy alone to radiation therapy plus darbepoetin alfa in the treatment of advanced head and neck cancer, trying to build on those preliminary studies I had shown you, indicating that you can improve survival and disease progression.

They corrected the hemoglobin to 14 g/dL to 15.5 g/dL, a very high target range, with the idea that an oxygenated tumor is going to respond to radiation therapy better than a nonoxygenated tumor. What they found, however, was that the 3-year locoregional control in subjects treated with the erythropoietin level—I am sorry, with darbepoetin alfa—to these levels was significantly worse than those who did not receive the erythropoietic agent, and it was statistically significantly different. And also, overall survival also favored those not treated with the erythropoietic agent, that is, treated with placebo, although this did not reach statistical significance so that is one of the clinical trials that you have been reading about.

The second clinical trial I want to show you is one that was published last winter from Singh in CKD, and these would be patients who are predialysis with the hypothesis that if you normalized hemoglobin, the high hemoglobin group, the target being 13.5 g/dL and the low hemoglobin group, the target being 11.3 g/dL, so that is a 2-g difference, a clinically important difference, the hypothesis being, if you could normalize the hemoglobin in predialysis patients, somehow you would improve survival. They found the opposite. And this is the published data. The probability of a composite event being either mortality or hospitalization because of heart failure, myocardial infarction, cerebrovascular accident, serious cardiovascular complications, or death. A higher rate in the high hemoglobin group, 25% to 30% over more than 3 years, compared to about 20% in the low hemoglobin group.

In summary, my view on this is that anemia does remain an underrecognized and undertreated disease. I think in patients with chronic kidney disease, long-term anemia is a risk factor for left ventricular hypertrophy and mortality, but so far to date, the idea that you normalize hemoglobins is not a good idea, so stick with the National Kidney Foundation guidelines of 11 g/dL to 12 g/dL, and these may be revised going forward. We know that anemia does increase the cost of delivering health care in the chronic inflammatory state across the board, and most importantly, effective treatment of anemia should not be solely based on achieving an optimal hemoglobin level, but primarily as a blood transfusion sparing effect. Thank you very much.