Chapter 08: A Research Controversy: Treating Hemorrhage in Childhood Leukemia


Chapter 08: A Research Controversy: Treating Hemorrhage in Childhood Leukemia



Media is loading


In this chapter, Dr. Freireich talks about the controversies surrounding treating hemorrhaging in pediatric leukemia patients.



Publication Date



The Making Cancer History® Voices Oral History Collection, The University of Texas MD Anderson Cancer Center


Houston, Texas

Topics Covered

The Researcher; Overview; Definitions, Explanations, Translations; Discovery, Creativity and Innovation; Discovery and Success; On Research and Researchers; Professional Practice; Understanding Cancer, the History of Science, Cancer Research; The History of Health Care, Patient Care; Technology and R&D; Patients, Treatment, Survivors

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.


History of Science, Technology, and Medicine | Oncology | Oral History


Emil J Freireich, MD

As I said, Brecher and Cronkite had proved that platelets don't work. They had irradiated animals. If you irradiate an animal, he has no platelets. He doesn't bleed. They don't need platelets. You have to have something else happen. If you give them a little bit of heparin, an anticoagulant, they bleed like a stuck pig, so the cause of bleeding is some anticoagulant. It's not due to the platelets. Platelets are permissive. But I had discovered something. So I went to Dr. Zubrod, and I said, "I have this idea. If I can elevate their platelet count, I think we can stop the bleeding."

I had this child whose parents were there. How can you give them platelets? Well, platelets are very fragile, and they don't store. We didn't know how to store them. The only platelets I could get would be fresh platelets. So I got a brilliant idea. I said, "What if we did what Dr. Duke did in 1910, that is, an exchange transfusion? What if I got some adults, and I took 50 cc of blood with normal platelet count out, and I took 50 cc out of the child and put the normal blood in, and I kept doing that?" I still have the graph from the first child I did. It never was published.

Lesley Brunet, MA

It wasn't accepted for publication? It was too controversial?

Emil J Freireich, MD

Couldn't get out of the NIH. So I did this study. We did an exchange transfusion. We got volunteer donors from the church, and each donor donated 500 cc of blood, then 50 cc in type. Then we got another donor and just kept going. Normal platelet count is about 200,000 per microliter. The child had none.

It turns out, by exchange transfusion, the theory has been worked out. If you do 2 total blood volume exchanges, you will end up with 90 percent replacement of everything. To get up to 200,000, we had to do 2 blood volume changes. He's a little child. He's 1 square meter, so he has about 2,500 cc. So we had to do 5 liters. We had to have 10 donors, 500 cc each. No bleeding.

Lesley Brunet, MA

For how long?

Emil J Freireich, MD

For 5 days. And we counted the platelets. When the platelets got too low, the bleeding started again. That told you something. The most important factor was the number of platelets in their blood. How can I prove that? Brecher and Cronkite went to the scientific director of the Cancer Institute, not Dr. Zubrod, and said, "Freireich is doing experiments which are dangerous." Dr. Mider called me to his office, so help me God, and he said, "Freireich, Dr. Brecher and Dr. Schmidt tell me you're doing these experiments with platelet transfusions, and they think it's totally reckless. You have to cease and desist or you're going to be fired." I went to Dr. Zubrod, and I said, "Dr. Zubrod, what should I do? You told me to cure hemorrhage. I'm curing hemorrhage." Dr. Zubrod said, "The only thing to do is to make it explicit."

So he called a meeting—grand rounds on the top floor of the Clinical Center. Brecher and Cronkite were there; Dr. Mider was in the room. Schmidt presented the data, and I presented my data. At the end, Brecher said, "Heretofore, we will not allow Dr. Freireich to collect fresh blood for his patients because it's bad for them." Dr. Zubrod stood up and said, "As long as I have the responsibility to be director of the Cancer Institute, if my physicians order fresh blood, you will have to deliver it." It was a tremendous confrontation.

There's a little more background to it, because when I made this observation we said, "Okay. All we need is platelets." Where do you get platelets? Well, we can collect platelets from people and separate them, but we knew that they didn't survive in vitro. At room temperature or at cold temperature, at any temperature, the platelets dissolve. By the time blood has been stored under blood bank conditions for 48 hours, for red cells, you can't find any platelets; they're all dead.

We said, "There's only one way to approach this. We have to do a randomized prospective trial. We have to do it in a circumstance where the data are objective and quantitative to resolve the controversy. Here's what we're going to do. When a child is bleeding, we're going to call the blood bank, and the blood bank is going to decide by random allocation from a statistician whether that child gets fresh blood drawn from donors or blood that's been in the blood bank for more than 48 hours. We'll see if the platelets have any effect on hemorrhage."

Lesley Brunet, MA

But you knew that they would.

Emil J Freireich, MD

But they knew that it wouldn't. Remember, the opposition is more certain than I am. First of all, Brecher has been doing this for 25 years. The gospel is known. There's a body of data there, 4000 references in literature. All there is is me, dumb Freireich. He doesn't know up from down.

Lesley Brunet, MA

What about the data?

Emil J Freireich, MD

That's an interesting comment, "the data," because I still deal with this today. The data doesn't speak for itself. Data doesn't exist in the world; it exists in one's brain. When people look at data, they come out with different things. They saw the data, but they didn't believe what we had done. They said the exchange transfusion worked because I removed the anticoagulant. But I knew it was the platelets, because I had taken platelets from a normal donor in vitro and mixed it with the patient's blood in a test tube, and it corrected it. So there was no anticoagulant. But they had another hypothesis based on their own experiments. So everybody knew the answer. Everybody who agreed to this, Brecher and Cronkite, knew that it would be negative. They knew because they had the knowledge. I didn't know anything. I'm a dummkopf from nowhere. I had never worked in platelets or blood. I didn't have any publications. This is a new field for me. They agreed to this experiment. It was double-blind. The blood banker didn't know, the doctor didn't know, and the patient didn't know. We didn't have any stopping rules in those days. Statistics weren't too sophisticated. We treated a number of children, and some number came along.

One day Tom Frei said, "Freireich, we should see how it's going." It was obvious to me, taking care of the children, that in some children, after the blood, they stopped, and in other children, after the blood, they didn't. There was heterogeneity in outcome. Maybe we could see what it did. We had a public meeting. Dr. Schmidt, Dr. Brecher, who knows it's not platelets, Dr. Frei, me, and our statistician, Martin Schneiderman. We're going to break the code. So I go to the board and start with patient number one. "Did this patient stop bleeding or not?" Well, the way I assessed bleeding was not only looking at blood, but we measured the concentration of hemoglobin in urine. We measured the plasma hemoglobin, the rate of fall, so we had quantitative data. Patient number one received transfusion on this day. Hemorrhage change, yes or no? Yes. And we went on for several more patients. A hundred percent of those that I felt had changed had gotten fresh blood. Ninety-five percent of those who I said didn't change had gotten bank blood, because some of the children that got fresh blood didn't respond.

This is a serious problem now. Everybody left the meeting depressed. Two days passed, and I was called to Dr. Zubrod's office. "Dr. Brecher has pointed out that this study is null and void." "Why?" "Because you interviewed the families and asked them whether they donated blood or not, and most of the fresh blood donors were family members who were in the hospital. You knew in advance who was fresh and who was banked, and you cheated."

Lesley Brunet, MA

How does knowing in advance affect the results?

Emil J Freireich, MD

I could say that the ones who stopped bleeding had fresh blood, because I knew they got fresh blood. It wasn't blind. They were saying I knew they got fresh blood, and I said they stopped bleeding. The whole study is null and void. I'm a liar and a cheat, and I'm about to be fired again. How do you react if someone calls you a liar? I'm working my head off 20 hours a day, and I'm cheating?

I wrote the paper, and Dr. Frei reviewed it. He was my friend, and he believed me. Dr. Zubrod reviewed the paper, but he was in a tough position. Should we accept it or not? His career was at stake. If it was me alone, he would never have approved it, but Tom Frei put himself in between, and he trusted Tom. He couldn't turn his back on Frei and Freireich, so he let it go. It was published in the New England Journal of Medicine. It's a classic, of course. It's been confirmed a million times over. And Brecher put his name on the paper. We had a series of meetings after that with Brecher, and we said, "Come on. Let's face it."

Brecher was very famous because he invented the phase microscope method for counting platelets. Prior to that, platelets were counted by an indirect method that was invented by Dr. Dameshek. The indirect method is, you make a blood smear, and you count the number of platelets relevant to the number of red cells. You count the red cells objectively, and that gives you the platelet count. It turns out that tremendously overestimates the number of platelets, because the platelets adhere to glass and the red cells don't, but that wasn't known at the time. So Brecher invented, with Cronkite, the Brecher-Cronkite method for counting platelets directly, in a hemocytometer chamber. He used a phase-contrast microscope. If you change the light phase, then things become doubly refringent when they're in the out-of-phase light. You could see the platelets, so then you could count them objectively, and we knew the real platelet count.

I went down to Brecher's lab, and I learned how to do this. In the children who got the transfusions, I did a platelet count on every child. They accused me of lying, because I interviewed the family. Now, was there any way I could lie when I counted the platelets? They couldn't figure that one out. Well, it turned out that the difference between the children who got fresh blood and the ones that got bank blood was 5000 platelets per microliter, measured 1 hour posttransfusion. The normal is 200,000. No one could believe that 5000 per microliter could make a difference. But Brecher couldn't do anything about it. I had used his method. It was done double blind. There was no way for me to know which child got fresher blood before I knew they were bleeding or not. I had a technician who counted the platelets, so I know they're objective. That's why I put his name on the paper. He never believed that it was true, that fresh blood was better, but he believed the platelet counts. Then we did a brilliant study. I had a fellow whose name was Larry Gaydos, and we said, "Look, if the platelet count determines the risk of hemorrhage, let's go through the charts of 130 children that we have treated, examine the nurses' notes, decide on which days they had gross hemorrhage or minor hemorrhage, record that for every day, and then we'll record their platelet counts by the Brecher method." That paper is a citation classic. It's in the 100 most-cited papers in the literature. It was a retrospective review. We showed that there was a direct relationship between the platelet count and the hemorrhage. Aha! It's not an anticoagulant; it's the platelets.

So we said, "Wait a minute. If one transfusion of blood will give you 5000 per microliter, what would 2 units of blood give you?" It would give you 10,000, maybe. If 1 unit causes hemorrhage arrest for 3 to 5 days, how can we get more platelets? Well, we can give them 2 units of whole blood. But these kids only have 2,500 cc. They only have 5 units. They're already volume expanded. If you give them 2 units, they're going to die of a stroke immediately. We have to figure out a way to give them platelets.

I went back to the library. To make a long story short, we invented plateletpheresis. I was working with a young guy who was a fellow in the blood bank. His name is Allan Kliman. You see, the plastic bags had just been introduced into medicine. Fenwal Laboratories had introduced plastic into blood processing. Prior to 1955, all blood was collected in glass bottles with rubber tubes and steel needles, and it turns out that those things are all incompatible with platelets. By the time the blood gets to the bottle, 90 percent are removed by the rubber and the glass. But it's not bad for red cells. Red cells do okay. They don't stick to anything. Fenwal introduced the plastic line into the blood bank. You could collect blood in a plastic bag, and you could ship it, store it in the refrigerator, and it takes less room.

When you studied platelets, we knew that the surface was what determined whether the platelet lived or died. The platelet is designed to recognize a surface that is wettable. Platelets and plastic did very well. We could get a plastic bag, collect a unit of blood, spin it in a centrifuge, take the platelet-rich plasma out, give the red cells back to the donor, collect a second unit of blood, and do the same thing. We'd have 2 units of plasma. Plasma has the great virtue that the space for plasma is at least twice the space for red cells. So we can give 500 cc of plasma and get 2 units of blood from each donor and give it to a child. You just had to collect a lot of platelets, count them, and transfuse them.

Sure enough, when we finally did all the work, it turned out that you got about 12,000 increments per 1011 platelets. That's the amount in 1 unit of blood. If you do the plasmapheresis perfectly, instead of 10,000, you actually get 24,000. And 24,000, it turns out, is outside of the hemorrhage range. By the time you get over 20,000, the likelihood of bleeding is very low. We said, "Let's give these kids 2 units of platelet-rich plasma twice a week." The lifespan had been measured when I was working in Boston. Frank Ebaugh, who discovered the chromium method for the turnover of platelets, was working there, so we knew the half-life was 5 days. "Let's do 2 units twice a week." We did that in a 100 children and wrote a paper. It was rejected by every journal. We finally got it published in the Annals of Internal Medicine, and the rest is history.

Emil J Freireich, MD

Interview Session Two: 30 July 30 2001

Conditions Governing Access


Chapter 08: A Research Controversy: Treating Hemorrhage in Childhood Leukemia