The Pill That Could

Chapter 6: First in Human

The Oregon Health Sciences University Hospital sits on a hill in southwest Portland, in a complex of brick and concrete buildings that look out over the Willamette River and the Cascade Range beyond. The hematology clinic where Brian Druker sees patients is on one of the upper floors. Most days, sitting in the waiting room, you can see Mount Hood through the windows on a clear morning, and on a less clear morning, which is most mornings, you can see the lower edges of the cloud bank that the building sometimes seems to be inside.

On the morning of June 22, 1998, in an examination room down the hall from the waiting area, the first patient enrolled in the phase 1 trial of STI-571 swallowed a single capsule of the drug with a paper cup of water. He had failed interferon. He had refused or was ineligible for transplant. His CML was, by any reasonable measure, going to kill him on the timeline this kind of CML kills people on. The capsule he had just swallowed contained twenty-five milligrams of the molecule. Twenty-five milligrams was the starting dose, chosen — as starting doses in phase 1 trials always are — to be lower than the dose anyone expected to do anything. The point of the morning was to find out whether anything bad happened. The point of the trial was to find out, by climbing slowly upward in dose, where the bad things would start.

Nobody in the room expected the drug, at this dose, to do anything good.

It was a phase 1 trial. Phase 1 trials are not, in conventional cancer drug development, where benefit shows up. They are where toxicity shows up. The patients in them are typically people who have run out of other options, who consent to taking a drug at doses too low to help them in order to give the field information about whether the drug is dangerous. The benefit, if there is one, comes in later trials at higher doses. This is what everyone running and enrolling in the trial understood.

Druker, watching the patient leave the exam room, did not know what would happen next. Nobody did. The animal data suggested that the drug should be tolerable in humans. The cell data suggested that, at doses considerably higher than twenty-five milligrams, the drug should kill CML cells. Whether either of those would be true in an actual person was the question the trial existed to answer.

* * *

Phase 1 trials in cancer follow a standard design. You enroll a small number of patients — typically three — at the lowest planned dose. You observe them for some weeks; you watch their blood counts, their liver function, their kidney function, their general condition. If the dose is well tolerated, you escalate. The next three patients get a higher dose. If that's tolerated, you escalate again. You keep doing this, three patients at a time, until you start seeing dose-limiting toxicity — side effects severe enough that you wouldn't want to push higher. The dose just below that becomes the maximum tolerated dose, and that is the dose carried forward to phase 2.

The escalation steps in the STI-571 trial were not small. The starting dose was 25 milligrams once a day. The next planned dose was 50. After that, 85; then 140; then 200; then 300; then 400; then 500; then higher if needed. The plan covered a range of more than twentyfold, on the expectation that somewhere in that range you would either find toxicity or find efficacy or both, and on the further expectation — derived from animal data — that interesting drug levels in the bloodstream would not be reached until the upper portion of the range.

The first three patients took 25 milligrams a day. Druker and his team watched them. Nothing alarming happened. After several weeks, with no toxicity to speak of, the trial moved up to 50 milligrams. Three more patients enrolled. Nothing alarming happened. The trial moved to 85.

It is somewhere in here — sometime between 85 and 140 milligrams, sometime in the late summer of 1998 — that something started to happen.

* * *

CML is unusual among cancers in that you can monitor it, week by week, with a blood test. The disease is the bone marrow producing too many white blood cells, and the white blood cells circulate in the blood, and you can count them. A normal white blood count is in the range of four to ten thousand cells per microliter. A patient with untreated CML often presents with counts in the hundreds of thousands. The count is the disease, in a fairly direct sense; if the count comes down, the disease is getting better, and if the count keeps coming down to normal, the disease is in some kind of remission.

There were patients in the trial whose counts started to come down.

Not all of them. The patients on the lowest doses, by and large, didn't respond. Their counts stayed elevated, more or less in proportion to where they had been before they enrolled. But as the dose escalated, more of the patients in each cohort began to show falls in their counts. The drops were slow, over weeks; this was not the dramatic same-day collapse of cell counts you sometimes see with cytotoxic chemotherapy. It was steadier than that. The white cells came down. The platelets, which had been elevated in some patients, came down too. The spleen, in patients who had had enlarged spleens, started to shrink. People reported feeling better.

All 13 chapters of The Pill That Could · Library · All courses