University of Minnesota

Frederick Goetz

Year: August 24th, 2001
Interviewed by: Blackburn, Henry


Frederick Goetz, MD, diabetologist at the U. of Minnesota Medical School, was a member of the Steering Committee of the University Group Diabetes Program (UGDP) and centrally involved in the trial and in the turbulent aftermath, criticism, audits, and vindication of its process. The clinical trial, which showed no advantage to control of glycemia or to oral medications vs. diet and insulin in adult-onset diabetics, was a shock that the community of diabetoligists and the drug industry tolerated poorly. This interview contains much of the detail and drama in that unique, pioneering, multi-center clinical trial.


HB: I’m interested in what you remember about the Medical Tribune and our responses to it and Klimt’s role. Al Feinstein’s ….

FG: Al Feinstein is the one who actually comes to mind first of all ….

HB: They accused us of fraud here in reading UGDP cardiograms. So, whatever you feel like talking about. Perhaps starting with your main contributions to the study as you see it [and its origins].

FG: Someone at NIH fairly high up, whatever component at NIH at that time had to do with diabetes, had been given by Congress I presume, probably on some legislator’s personal hunt for something to solve a problem, had dropped a special pot labeled “diabetes” into NIH. Apart from the individual research grants, the proposition was that this was to be prospectively worked out. The NIH person who’s name I’m sure is clear in the records, there were some very nice people, very relaxed. They went to Max Miller and asked him what they should do about this new money. Max got a few people together and I think Chris Klimt was at Hopkins finishing his own doctoral degree, I’m not quite sure. It may have been the NIH people that got him and Max together. Maybe half a dozen people got together somewhere and I remember somehow sitting in a pleasant cocktail lounge and talking about what could be done by way of natural history of diabetes. This was Max’s first thought and he certainly was very important in developing the whole concept. Somewhere in the back of his mind, I think Max had a kind of internal drive to examine what he often referred to as the “Joslin Clinic” way of looking at diabetes. He was very troubled by this ….

HB: Which was rigid control or ….

FG: Yes, exactly. Just control. The oral tablets, I’m sure this is all familiar to you, were just coming onto the market. They had been approved by FDA and were on the shelves of the neighborhood pharmacies all over the country. But the idea of tight control was mostly an insulin-related phenomenon with about 40 years of experience empirically with treatment of diabetes, insulin release, discovered about 1922, I think. So something like 40 years. Well, Max thought that the Joslin people who were preaching in almost a messianic way – this tight control – had not a leg to stand on when it came to evidence supporting it.

HB: Long-term.

FG: That’s right. Long-term benefit in respect to complications. There was no doubt about cutting down mortality from ketoacidosis, nobody was talking about that. But survival, and prevention of serious complications, especially in the eye and increasingly in the kidney. Max felt almost driven to propose a scientific way of examining that question. The question of whether one should study insulin-requiring diabetes or the much commoner insulin independent diabetes, that was not very clear at the first meeting or two with Max. But Max first spoke of natural history, “We don’t know the course of diabetes, do we?” Except for Harvey Knolls who had a study in Cincinnati and wrote one of the few papers that didn’t take a stand one way or the other on control. You may remember Harvey’s study in which, I think the title when finally published was “The Course of Juvenile Diabetes Treated with an Unmeasured Diet,” the idea was that there wasn’t very much on natural history without attempting to regulate the matter of blood sugar control, simply observing it. But gradually there was a shift from the idea of simply another natural history study to possibly a clinical trial with the oral agents now coming on the market. The idea really seemed very attractive. The new agents were FDA approved, but we really were in the beginning of our knowledge of these, shouldn’t we consider a clinical trial? And that was my first formal clinical trial.

But the key features – randomization, incorporation of placebo, double-blind where possible, and the classical ideas – this is where I had my education about all of this. Of course, the appeal of studying the non-insulin-dependent diabetic was that you could ethically randomize to placebo and at least the immediate risks of allowing blood sugar to be increased were not visible. And the folks on the long-term aspect of it came up right away. I think that led then to what was really a strong protocol, with great care taken on the assignment of treatment, and the randomization process. There were several horror stories about previous clinical trials in the area of vascular disease. I think there was one on anticoagulants and pulmonary embolism, as I recall, in which the people actually doing the assignment to treatment or placebo, through a flaw in the way the randomization was carried out (such things as specifying that the emergency room patients coming in with pulmonary embolism, if they were admitted before twelve midnight they’d go on placebo). People would fudge all the time. So this was very tightly designed. This was impressive to me. The plan was tight as was the execution I can certainly vouch for from personal experience. There was no way to get bias from physician’s judgment. You couldn’t say, “I want this person to be on treatment A or B.” All that was most carefully worked out. Another good thing about it was that the definition of non-insulin-dependent diabetes was quite rigorously carried out, with a waiting period. As you may recall, the subjects being proposed for entry into the protocol had I think it was two months off insulin treatment. And if they didn’t make it to two months because of recurrent therapy for ketosis, they would be ineligible for the study. So that the definition of acceptable cases was good.

HB: The outside criticisms of selection, I think, were ill-conceived.

FG: That’s right. They were, as eventually was clear, due to some misunderstandings about blood sugar methodology. A few people entered into the study who weren’t really diabetic, didn’t meet criteria for the study. But those were a very small number of subjects. What in my own mind is still the most crucial part of this whole story, was the decision to drop Tolbutamide after, well, actually I don’t know how many years it was. It was late 1969. By the way, you might know this volume. It’s a monograph by Curt Meinert and he has a long discussion on the UDP as a case study in his text.

A dramatic meeting

FG: The first patient was entered in February 1961. The presentation of the Tolbutimide results at the American Diabetes Association was in June 1970. That meeting of the ADA, I don’t know if you happened to have attended that, was in St. Louis. The weather was hot and steamy and tempers were hot and steamy. The tension, really, was just fantastic!

HB: How did they know about it? Because there was some criticism that the information was inappropriately released.

FG: Lots of complaints. For some years I was invited by an ethicist at St. Thomas to come and talk about this study and it was a very good piece of decision-making to share with the students. The Declaration of Helsinki had then fairly recently promulgated the issues in an updated form, the key feature being that where physicians are involved as responsible persons for investigation, their one-to-one relationship with patients took precedence always over general benefits to society or the larger public, even though results for the good of mankind might be sacrificed by doing this. This is very clear and it made very good sense. When we looked at ourselves in the UGDP we (the investigators) realized that we were in the peculiar situation of being the only physicians or health-related individuals who were aware of the possibly serious risks from using the medication in the study. The rest of the world of medical care was blissfully unaware of this. What were our responsibilities, therefore? They were to our own subjects, who in good faith had entered this study with oral informed consent. What was in our minds were these Monte Carlo odds. Here we were with these data before us, we had no choice but to act. But the Peter Brigham and the Joslin Clinic people were saying, “You’re crazy, what are you doing? You’re saying that this drug which has been considered a boon to mankind, the first new treatment for diabetes in 40 years is a danger?” And this is where tempers rose and rose. Of course, a lot of this was released, I won’t say leaked, but it was released to the Wall Street Journal. [It turned out that early receipt of the article was a routine for the publication Diabetes, required for its printing, so it was not an unofficial leak from the study.]

HB: The basic conclusions of the study, you think, were justified at the time?

FG: Oh yes. I think that the decision to drop Tolbutamide was right.

HB: But the fact that nothing was any better than insulin, diet… The major conclusion was that insulin-variable dose wasn’t significantly better.

FG: That’s unfortunately the case. Now since then, there’s been the UK study which …

HB: Harry Keen’s study?

FG: It certainly grew out of Harry Keen. Robert Turner, I suppose was the chief organizer of that. But that was just reported in the last year or two. It’s a massive study on United Kingdom residents in a whole cluster of diabetes clinics ….

HB: Examining the issue of close control or ……

FG: It had many good features including randomization. It was not double-blind. It involved thousands of patients, maybe three times that of UGDP. Overall, the study shows a benefit including reduced cardiovascular events….

HB: Macro and micro complications were improved by good control in adult onset diabetes?

FG: That’s right. It’s generally favorable …..

HB: There have been a number of other studies on microvascular disease coordinated from here in Minnesota and all sorts of other studies, haven’t there?

FG: The most impressive, I would have to say, is the study of insulin-requiring diabetics. The so-called DCCT. That’s one that’s everyone’s happy about. The Joslin people and everybody else. When that study result was announced about 40 years ago at an ADA meeting, the Joslin Clinic people wore big buttons, “We told you so.”

HB: That’s a wonderful story! But that was insulin-dependent diabetes

FG: That was insulin dependent. That was a massive thing. Of course there were people, especially in Europe, who said you didn’t have to spend all that money. We always knew that it was treatable. But in general, I think people agreed that it was helpful. But that was insulin dependent. What’s loosely called the UK Study of Non-Insulin-Dependent Diabetes is not nearly as crisp and should you wish to dive into that I could, perhaps, mention one or two papers that are quite nice in summarizing it.

Back to Tolbutamide

HB: That actually led to very nasty things. That meeting you’re talking about was the Tolbutamide meeting?

FG: That was the Tolbutamide meeting?

HB: What do you recall about Feinstein’s role?

FG: He was the pet expert witness of the drug industry.

HB: He was their paid expert, shall we say?

FG: His tactics toward the principal investigators, including me, was to patronize us as lambs led to the slaughter by some sinister forces masking themselves as scientists.

HB: Who were these sinister forces?

FG: These were the statistical people. Chris and Company.

HB: Personified by Chris Klimt

FG: And saying how you really shouldn’t believe all these things that those statisticians tell you. Of course it implied that we were a bunch of dummies. Max Miller was terribly upset by that. So if anything made us feel more loyal to the study, obviously it was that. Feinstein wrote some very pertinent discussions of the issues. But he ended up, I think, making a wrong conclusion that the study ….

HB: He concluded that there was fraud. Or I think he suggested it.

FG: Yes, he implied dishonesty. When Chris Klimt was at the European Diabetic meeting, the German drug industry was very distressed. I never did go to some of those meetings, but there was someone to whom Chris Klimt was very grateful, a well-regarded European diabetologist who stood up and said, “Colleagues, forebear, forebear. This study has been carried out by the approved scientific methods. There is no reason to think they were not so done and let us be careful before condemning something.”

HB: That’s better than anybody in this country did. How is it that Chris brought so much trouble on himself (other than his personality)? What was his fight about releasing the data? That he won?

FG: That went on and on. I’m refreshing myself. It went all the way to the U.S. Supreme Court, which I had forgotten. There was a question of whether the Freedom of Information Act required investigators to release all data. And the Supreme Court decided 6-2 and said “No, they don’t have to do that.” Eventually, as you’re probably aware, the data were deposited in, I’ve forgotten, Ann Arbor or somewhere and are generally available. I gather there hasn’t been much of a demand, after the fact.

HB: What were the grounds for trying to get the data and grounds for protecting the data? Do you remember? The obvious grounds that “the people” paid for it, thus “the people” should have access to it. What were the arguments that Chris used and the Court used to justify protecting the raw data? Do you remember? I’ll look it up in the book. I’m glad to know about this source.

FG: Let me just give a quick look. Besides the chronology, Curt has a list of comments and rebuttals. Holbrook Selzer was a diabetologist in Dallas. I notice it because it has to do with ECG coding.

HB: I had better borrow that from you.

FG: It’s amazing. Holbrook complained that there were changes in the ECG coding procedures midway in the course of the study. I won’t take more of your time. But I think you’ll find a lot of interesting reading there.


HB: So the UGDP stimulated a great deal of thought and elevated the science of the study of diabetes, irrespective of any other contribution.

FG: Absolutely. And as time passed, I think it was given not its full due of respect, but a grudging respect. There have been, of course, in vitro animal studies that have tended to support UGDP as far as this question of mortality from a myocardial infarct. You may be fully aware of that, of animal studies indicating that rhythmicity of rabbit heart muscle in vitro is impaired by adding Tolbutamide or similar drugs. There have been studies of a tissue receptor for Tolbutamide which has implications for the nature of the oral agents of that group. It’s conceivable that a mechanism for excess mortality may exist by way of altering rhythmicity of the heart. I’m putting this in very general terms.

Full Transcript Access

Full transcripts of interviews may be made available to those engaged with original materials for scholarly studies by contacting us.