digital images to review them myself, slice by slice. Her brain looked normal, but I knew from Irene's symptoms that it was dying on one side. It was simply too early for a CT to detect the damage. By tomorrow, I knew that half of Irene's brain would be black and swollen. I also knew that the swelling would continue to progress, peaking in about 48 hours. By the third day, without treatment, there was a good chance she would die.

"CT has ruled out a brain tumor or hemorrhage," I said. "The blood tests look good. I have her consent to treat." I expelled a long breath. "Let's do the tPA and let's do it now!" I heard the determination and urgency in my voice. I just hoped my bravado was matched by sound judgment.

There was no one else to speak with. Irene lived alone, and her family hadn't yet been located. In any case, we had no time to lose. By now, the Dilantin had come up from the pharmacy and the nurses were setting up the tubing to pump the medicine into the vein in Irene's left arm. Minutes later, the tPA arrived. I saw that Irene was just coming to from her seizure. It was now or never.

Slowly and carefully, I pushed the needle into her IV port and depressed the plunger on a syringe full of tPA. I winced and braced myself for disaster, aware I was courting a risk of death for my patient greater than undergoing open-heart surgery. I pictured the tPA acting like Drano in a rusty pipe. It would either break open the clogged pipe or blow a hole in it, causing massive brain hemorrhage and death. I prayed for the former.

But bending over Irene as I watched her vital signs closely, I felt oddly peaceful. I could still hear Irene's confident whisper in my ear: "Give me the drug." And I saw again the look in her eyes—a look of serenity that could only spring from faith. I was humbled by the knowledge that Irene's leap of faith was in me, a total stranger making rapid decisions that

"Give her two milligrams of Ativan stat and let's get one gram of Dilantin rolling," I directed. Using both drugs as a one-two punch is usually very effective. Ativan stops convulsions immediately, while Dilantin keeps them from recurring.

could transform and save her life—or end it.

This, I knew, was really the essence of medicine—our best guess, based on the available research, our own experience, and what we know about an individual patient. It is no more, no less. Choosing a particular treatment requires faith— sometimes enormous, go-for-broke faith— on the part of doctor and patient alike.

We moved Irene to the intensive care unit, where I ordered an ultrasound of the arteries in her brain. The test, called transcranial Doppler, can often detect immediately how successfully tPA is opening a clogged blood vessel in the brain.

I continued to watch Irene closely. She appeared groggy and tired, but no better or worse than when I had first seen her.

When the Doppler machine arrived, I took over the controls: I wanted to see the

results myself. Then a fairly new test, the Doppler shoots a focused, low-frequency wave of ultrasound through the skull and into the blocked artery, allowing the doctor to determine on the spot whether treatment is making a difference. Within moments, I located the problem artery on the computer screen. I could see that the vessel had become unblocked and that blood, and thus oxygen, were already flowing back into Irene's damaged right brain. The tPA was working!

The night went well for Irene. She tolerated the drug and the fluids we were giving her, and when I saw her the next morning she appeared much stronger. She could even lift her left arm and leg off the bed. But she no longer recognized me and appeared agitated and disoriented.

A follow-up CT scan confirmed that Irene had indeed suffered a massive stroke-in-the-making. Her right brain was still somewhat swollen, but within the areas of dead tissue were islands of normal, living tissue. Much of her brain had been saved by the reflow of blood. The miraculous but dangerous tPA had done its job!

By day four, the swelling had gone down and Irene's mental sharpness was much improved. But when I saw her on rounds that morning, she looked at me accusingly. "I thought you said the drug you gave me was going to make me all better," she said. "But it's not so!" Her voice was tight with disappointment. By now, she had some feeling and function back on her left side, but her left arm and leg remained weak and she couldn't yet walk without assistance. "I can't function this way!" she told me angrily.

"I know, Irene," I said sympathetically. "It's too soon to judge how complete your recovery will be."

As she looked back at me, at once outraged and vulnerable, I felt myself sink into self-doubt. Irene had been in the hospital for almost a week. She was old. She hadn't been given the drug till almost 100 minutes out from her symptoms.