(FORTUNE Magazine) – Few media circuses surpass the one that swept through the University of Utah Medical Center in 1982 after it announced the implantation of the Jarvik-7 artificial heart in Barney Clark. Millions tuned in; journalists who stampeded to cover the story got so pumped up that some reportedly cornered hospital staffers and offered bribes for scoops. Inventor Robert Jarvik, 36, seemed about to walk across the Great Salt Lake in a burst of boyish ebullience. Then the Lazarus aura faded: Clark, 61, sank in an agony of complications and died on Day 112. The Jarvik-7 was implanted four more times to replace failing hearts, with similar results, before federal authorities halted the procedure.

In contrast, the media all but ignored the implantation this year on April 10 of a heart pump in Francis McKeon, 68, a retired computer programmer, at New York's Columbia-Presbyterian Medical Center. The national coverage consisted of a brief Associated Press story and a TV news spot. Yet the Grim Reaper threw a fit. For the first time, an American dying of heart failure would be going home to live an essentially normal life--potentially for years--with a mechanical device swishing blood through his body.

McKeon's death cheater is a left-ventricular assist device, or LVAD, made by Thermo Cardio-systems of Woburn, Massachusetts. It works in parallel with a moribund heart, shouldering most of its burden. Since 1986 the LVAD has been implanted in more than 700 people for up to 17 months as they've awaited human heart transplants. Surgeon Eric Rose, who put in McKeon's pump with colleague Mehmet C. Oz, says the procedure is now almost routine. But the results aren't: Supported by the bagel-size pump, formerly ashen cardiac patients have outperformed their doctors on treadmills and even shot hoops. Its quietly mounting successes presage a revolution. Within a decade artificial hearts could be beating in thousands of breasts, adding quality years.

If permanent artificial hearts are cardiology's moon shot, McKeon's LVAD is the equivalent of Gemini, the Apollo program's earth-orbiting precursor. The device incorporates a host of hard-won technological advances. Perhaps the most important are its "bio-compatible" materials, which have allowed LVADs to function without problems for well over a year in patients' bodies.

Columbia-Presbyterian's Dr. Oz estimates that 20% of patients with failing hearts, particularly those too old to receive transplants, may be good candidates for permanently implanted LVADs. The device in McKeon's lower chest connects to his heart's main pumping chamber, the left ventricle, via a short tube. Once enough blood flows through the tube to fill the LVAD's pump, sensors prompt its electric motor to rotate, pushing a pistonlike plate that propels the blood into the aorta, the main artery leading from the heart.

When McKeon heads home from the hospital in June, he will wear under his clothes a shoulder holster with two 1.5-pound rechargeable batteries that can power the pump for up to eight hours. The batteries are hooked to a wire leading through the abdominal wall. The only detectable sign that McKeon is wearing the device will be a faint swishing. "Sometimes I do a stage trick at medical meet- ings," says Dr. Oz. "I put an LVAD patient in the audience, and nobody notices he's wearing the device until I ask him to stand up."

Like most conjuring, Dr. Oz's trick was paid for in buckets of sweat and countless fumbles. No one knows the costs better than Victor Poirier, a 54-year-old mechanical engineer who led the development of Thermo Cardiosystems' LVAD and is now chief executive. After working 30 years on the project, he's thoroughly acquainted with the bitter taste of humble pie: "When we started in 1965, we didn't know anything about blood clotting and infections. The heart was just a pump. We were engineers and we knew how to make pumps."

It took 20 years of lab work before Poirier and his team were ready to implant its device in a human. They needed another decade of clinical trials before the FDA approved tests of long-term use. Forced back to the drawing board time after time, they once scrapped a major component after sinking ten years into it. After countless animal tests, they watched with dismay as the first few people hooked up to their prototype pump died on the operating table--surgeons initially used it only when patients were nearly dead. But they eventually showed they had what it took to mimic designs honed by three billion years of evolution.

The first requirement: Hire a dogged perfectionist like Poirier. Says David Gernes, an electrical engineer who joined the team in 1973: "Very often it was Vic's attitude that kept us going. He always believes there's got to be a way." Poirier proved it by spearheading more than a dozen patents. Among the most important was a pioneering way to make an artificial pump compatible with the body. The technique prevents life-threatening blood clots by encouraging cells to grow in a thin layer on the pump's inner surface.

But as Poirier and his crew labored in relative obscurity, Jarvik's team sprinted ahead to Promethean renown. A physician, Jarvik dismissed devices like Thermo's as the kind of Rube Goldberg contraption you'd expect from gear wonks. "Pure engineers have made very complicated pumps for artificial hearts," he told FORTUNE 15 years ago, before Clark's implantation. "My pump is very simple."

So simple that "it made my hair stand on end," recalls Poirier, a phlegmatic man with a walrus mustache and a vague air of long suffering. "For instance, it had aluminum parts. Aluminum dissolves in the body." An artificial heart, he adds, must beat 100,000 times a day for years, assailed by warm, salty blood, plastic-dissolving enzymes, and cells that try to clog it like algae in a stagnant pond.

Yet there was justice in Jarvik's jab. Poirier's employer at the time, Thermo Electron Corp., which later spun off the heart-pump company, had developed nuclear power sources for space probes in the 1960s. When awarded a National Institutes of Health grant to work on an artificial heart, Thermo's engineers figured that going nuclear was the only way to make a compact power source that would last for years--a strategy pushed by the late Edward Teller, a Thermo director and father of the H-bomb.

The result was like a prop from Dr. Strangelove: a pump driven by a miniature steam engine that drew heat from a lead-shielded kernel of plutonium. "It was the most complicated system you could think of," sighs Poirier. But it did work. Thermo even tested it in monkeys, then bred the animals to show that stray radiation wouldn't cause mutations. The idea was finally killed, mainly because of fears that terrorists would kidnap heart patients and use their plutonium to build bombs. Now Thermo's pumps are driven by either compressed air or an electric motor.

Poirier's team bogged down on other fronts. The stickiest problem was clotting--when implanted, most man-made materials instantly trigger the body's plug-the-dike routine. Patients who received the metal-and-plastic Jarvik-7 had to be kept on blood thinners to prevent clots, putting them at grave risk of fatal bleeding. They suffered strokes anyway, apparently because clots formed on pump surfaces, broke free, and lodged in their brains.

Thermo's team tried scores of materials and hundreds of designs, sometimes testing prototype pumps for years in saltwater tanks and in animals before discovering fatal flaws and starting over. Finally they found the right stuff for durability: titanium, a light, strong metal, for the pump's shell, and a polymer similar to the Lycra in garments for the lining. But blood clots continued to bedevil them until they figured out a way to co-opt the body's own metabolic machinery.

According to conventional wisdom, clots were best averted with ultraslick pump linings that would prevent blood from getting a toehold for coagulation. But some studies suggested cells in the blood might colonize a rough surface without clotting, forming a thin layer that would mimic the wondrously clot-resistant lining of arteries. Unfettered by medical orthodoxy, Thermo's engineers opted to pursue the rough approach while Jarvik and other rivals went slick.

At first, Thermo's bet seemed hopeless. Many of its prototype pump linings rapidly fell apart. When implanted in animals, they generated fibrous growths that could break free and kill. In desperation, the engineers invented a chopping tool to slice polymer strands into synthetic fuzz. Then they rigged up a kind of electric shotgun to fire the fuzz onto plastic targets in hopes of embedding it in patterns that would foster the desired cell layer. No luck. It took years to finally find a way: "pre-seeding" suedelike polymer surfaces with fast-growing cells just before implantation. The trick enabled the pumps to work for more than a year in animals without signs of clotting, paving the way for human tests.

Poirier credits his doggedness to his father, a French Canadian immigrant who settled in Lowell, Massachusetts, and continued to work as an upholsterer after a back injury left him partly paralyzed. In 1961 the younger Poirier, married at 19 and unable to afford college, went to work as a junior draftsman at Thermo. "It was an MIT spinoff with a lot of very bright people," he says. "I envied them. It made me want to be an engineer in the worst way." Badly enough, in fact, to slog through nine years of night school to get a bachelor's in engineering from unsnooty Northeastern University. Earning stripes at Thermo for pluck and ingenuity, he eventually was named to lead the LVAD project through the tricky clinical-testing phase.

While Poirier and his team were sweating the details in animal studies, Jarvik's setbacks caused added frustration. The Jarvik team had launched a $19 million public offering for its new company, Symbion. Later, when patients' ordeals on the Jarvik-7 caused a backlash, Thermo's pump was tarred too. Says Poirier, who soon after tried to raise money for clinical studies: "I'd get up at investor meetings and hear comments in the back like, 'Goddammit, not another artificial heart."' But after honing his speaking and business skills in an after-hours MBA program at Northeastern, he slowly eroded the skepticism with straight talk and promising data. Hammering on solvable problems helped distract him from one that wasn't: His teenage son died in 1982 after a two-year battle with brain cancer.

For years, Poirier went into surgery whenever his pump did. When a nonimplanted prototype was first used to support the heart of a patient near death, he watched in horror as it began to falter in the operating room. Quickly he figured out why: The pump was driven by a frigid gas, and the operating room's cool air was overloading the heat exchanger that prevented the device's valves from freezing. The surgery went on for hours; Poirier kept the pump alive by wrapping his hands around the ice-cold heat exchanger. But the patient died. Says O.H. Frazier, a surgeon who led clinical tests of the LVAD at the Texas Heart Institute in Houston: "We lost our first four patients on the pumps. From a medical standpoint, it was very tough. But when we took the pumps out, they always looked good."

Surgeons eventually began using the devices earlier--before widespread organ failure--and got promising results. By the late Eighties, Thermo's LVAD was regularly used in clinical tests as a so-called bridge-to-transplant. When the FDA approved it for that purpose in 1994, it became the first implanted heart pump widely available in the U.S. The agency also recently gave Thermo a green light for clinical trials of long-term take-home use in selected patients with no alternative--Francis McKeon is the first one. Doctors are eager to widen the use of LVADs because, in several cases, mortally ill patients who received them have made stunning recoveries. Given a chance to rest by the devices, their hearts regained so much pumping power that the LVADs could be removed.

Thermo hopes to sell at least 700 LVADs this year, at about $50,000 each. Health insurers aren't balking at the bill: Besides sparing lives, the pumps keep patients out of costly intensive care. But Thermo's edge may not last. Its main rival, Baxter Healthcare's Novacor unit, is seeking FDA approval for an LVAD it already sells commercially in Europe; seven patients in Europe and Japan now have the device permanently implanted. And the NIH recently launched a $27 million five-year program to develop improved LVADs for permanent use. The money is earmarked for six teams, including ones at Nimbus, Abiomed, Transicoil, and Whalen Biomedical. The NIH also is sponsoring three teams trying to develop total artificial hearts.

Thermo itself is off the NIH's fiscal life support as it mulls improved LVADs. Its profit more than tripled last year, to $7 million, on $21 million of sales, and its share price has risen more than 3,000%, to around $54, since its 1989 initial public offering. One prospect that may have investors excited: analysts' projections that the market for LVADs could ultimately grow to $3 billion a year.

But Poirier is still hunkered down--the only festive note in his spartan office is a spray of plastic flowers. Says he: "I'm sure we'll see lots of problems moving to long-term use. There will be strokes and infections. There will be malfunctions and pumps that break. We might wind up in court. But that's another obstacle I think we'll overcome."