The Heart of the Matter

The heart. Often considered the center of emotion, the embodiment of the soul, the keeper of the spirit, it is seen by many as the source of courage, intuition and affection. It is the symbol of love. In reality, the human heart is a fist-sized mass of muscle connected to the rest of the body by a network of blood vessels more than 60,000 miles long. Running non-stop from birth until death, it will beat more than 2.5 billion times and pump more than 1 million barrels of blood.

Unless it stops.

Heart disease is the number one killer in the United States, claiming nearly one million lives per year. Since the 1950s, researchers have worked endlessly to develop artificial replacements and assist devices to stem this tide of failing hearts.

One of the preeminent researchers in the field of artificial heart development is Dr. Robert Jarvik, who has been heavily involved since the early 70s. He is probably most well known for his work on the Jarvik 7, the first permanent total artificial heart to be implanted in a human.

On a snowy night in December 1982, a Jarvik 7 total artificial heart was implanted into the chest of a 61-year-old Seattle dentist named Barney Clark. The world looked on as a team of surgeons removed his failing heart – so damaged that it stopped pumping during surgery – and replaced it with a mechanical assemblage of polyurethane and titanium. Connected by a pair of hoses through his abdomen to an external pneumatic pump, the Jarvik 7 kept the elixir of life coursing through Clark’s body, allowing him to survive.


Prior to surgery, Clark had literally been on the edge of death. He could not get out of bed and had to stay in a darkened room. Doctors even restricted visits by his wife – the excitement was too much for his failing heart. Medically not a candidate for organ transplant, Clark’s only alternatives were the Jarvik 7 . . . or death.

Following the implantation of the artificial heart, Clark was able to sit up, see his wife and eventually resume some semblance of functionality. Though constrained to the hospital, he was able to get up and around, and at one point was even able to practice putting a few golf balls.

Despite the success of the implant and the valiant efforts of his doctors, Clark eventually succumbed to complications and died. Though he only survived 112 days with his replacement heart, it was 112 days of life he otherwise would not have had.

The Jarvik 7 was the culmination of many years of work by a multitude of people. Chief among them was Dr. Jarvik, who spent more than a decade of his life modifying and perfecting the design to bring it to clinical trial.

But even before the clinical trials of the Jarvik 7, Dr. Jarvik was looking for a battery-powered portable system. Today, he continues his quest for a reliable, practical and forgettable artificial heart.


“If the artificial heart is ever to achieve its objective,” Jarvik stated in a 1981 Scientific American article, “it must be more than a pump. It must also be more than functional, reliable and dependable. It must be forgettable.” In other words, “It has to be so good that the patient goes about their daily life and, most of the time, doesn’t think about the fact that they’ve got an artificial heart.”

Jarvik’s latest effort at a forgettable artificial heart is the Jarvik 2000. Unlike the Jarvik 7, which was a total artificial heart and required removal of the natural heart, the Jarvik 2000 is a left ventricular assist device, or LVAD, which leaves the natural heart intact and acts as a booster pump. About the size of a C-cell battery, the Jarvik 2000 is actually placed inside the left ventricle of the heart, making it an intraventricular device.

“The left side does 80% of the mechanical work, or more,” Jarvik explained. “And the number of patients that primarily need left support only, rather than bilateral support, is probably about 75%. The estimates – year after year – keep indicating that about three-fourths of the people will only need a left-sided pump. That’s why we’re starting with that.”

Also, research has shown that it is better to leave the diseased heart in place, rather than remove it and put in a total artificial heart. “There’s a lot of evidence now that when you support sick hearts they improve,” Jarvik said. “If you boost the output of the natural heart and keep it working, you have the advantage that the heart still regulates the body’s needs and the amount of blood flow the patient gets.”

One of the issues that evolved from the clinical trials of the Jarvik 7 was that, in a number of patients, there would be infections around the outside of the artificial heart. “That was a hard thing to prevent,” Jarvik said. “In fact, with many types of heart-assist devices, there’s a very high incidence of infection.

“I invented the intraventricular heart to prevent infection,” he continued. “By making an artificial heart that is so small it can actually be implanted inside the natural heart, you have the advantage that the natural heart is there to support and protect it. But also, now the artificial device is surrounded by blood, and blood has antibodies and white cells and all the things that fight infection.”

Making a pump that would fit inside the left ventricle required a design that was very small and very efficient. The solution was a small axial-flow pump, where the blood flows through the pump and parallel to the axis of rotation, like in a turbine. Jarvik began working with axial-flow pumps and the battery-powered concept back in 1975 as a means to replace the external compressed-air system of the Jarvik 7. The idea was to use a miniature axial-flow pump to pump hydraulic fluid which, in turn, pushed the diaphragm of the artificial heart. “This would give you an electrically-powered, portable, very practical kind of device,” Jarvik said.

“It’s that miniature rotary pump technology that has evolved and developed into the Jarvik 2000 heart. The difference, of course, is that the Jarvik 2000 heart is a miniature axial-flow pump that pumps blood directly, so we’ve eliminated all the diaphragms and the valves and the size and complexity of the Jarvik 7.”

According to Jarvik, “The key to making a really miniaturized and reliable axial-flow pump is precision machining. The design has to be right, but we also need some fairly elaborate three-dimensional blade shapes, like turbo machinery typically has. We need precision, blood immersed bearings, and high-precision alignment of the pump parts that hold the bearings and rotor in order to get hydrodynamic support on the bearing film for a really reliable device.”

From the start, Dr. Jarvik has done his own designs and his own machining. “I got involved in machining pretty heavily when I started out in Utah,” he commented. “I would always machine most everything for the prototype models of what I was doing. I have a lot of experience in machining.”

Jarvik still does his own design work, but most of the machining falls into the hands of Michael Morrow at Jarvik Heart, Inc., in Manhattan, and Walter Wood at Transicoil Inc. in Norristown, Pennsylvania.

Michael Morrow is the research technician/machinist at Jarvik Heart, a self-contained, development-type machine shop located on the 15th floor of a Manhattan office building. Walter Wood is a manufacturing engineer at Transicoil, a leader in precision motion control, actuation systems, pilot and interface products for more than half a century.

Transicoil’s involvement stemmed from their work with fractional horsepower brushless DC motors. “We use miniature electric motors in these types of pumps,” explains Jarvik. “At first, Transicoil was building special little DC motors for the earlier model axial-flow pumps. Now they are the prime contractor on the NIH (National Institutes of Health) contract for an innovative ventricular assist system using the intraventricular pump.”

The NIH currently funds three major blood pump development programs that use rotary pumps. They are: the Jarvik 2000 at Transicoil; another axial-flow pump from a company called Nimbus; and a centrifugal pump from the Cleveland Clinic. Additional work is also under way in Europe and Japan.

“What we’re trying to do is prove the feasibility of this design, and demonstrate its effectiveness in humans,” states Jarvik, “and then have it grow from there.”

At present, work is under way on several different versions of the Jarvik 2000 axial-flow blood pump. There is the NIH-sponsored work at Transicoil, which is designed for long-term use and features an implanted power supply and redundant electronics systems. There is a version which has external batteries and electronics and brings power into the unit through a skull-mounted pedestal. And there is a plastic molded version under development by US Surgical Corporation for temporary use during surgery as an improvement over the heart-lung machine.

The key to making a reliable axial-flow heart pump is precision machining. From the outset Dr. Jarvik has relied on machine tools from Haas Automation for his research. In the early days he performed much of his machining on a retrofit mill with a Haas 4th-axis rotary unit. Today, he dedicates a pair of Haas VMCs with 4th- and 5th-axis rotary tables to his research: a VF-2 with a TRT-160 tilting rotary table at Transicoil, and a VF-1 with a T5C tilting 5C indexer at Jarvik Heart.

The pump blades and housing of the Jarvik 2000 are machined entirely out of titanium because of that material’s bio-compatibility, light weight and strength. It’s this last characteristic, however, that makes the components of the titanium heart difficult to machine.

Michael Morrow explains: “Titanium is really a difficult material to cut, and it can break down a tool quickly, just because of the nature of the material. You need a good, rigid spindle, and a good, rigid table. Of course we’re not cutting large pieces of titanium,” he continued, “but the tolerance we have to hold is within two-tenths (two ten-thousandths of an inch). Your spindle has to be very rigid and run very true; and the table has to have very little backlash, and you have to be able to compensate for the backlash very precisely. The Haas has really handled that very well.”

Walter Wood added, “Titanium is such a poor conductor of heat that if you get a little too aggressive with cutting, you’ll either burn up the tool or leave a lot of galling on the part. And, some of the tools we use are pretty small in diameter, so we can’t push them too hard.”

The Jarvik 2000 is a very compact unit – about 1" in diameter by 2.5" long – so the individual components are quite small. The parts are also quite precise and require a lot of complex multi-axis machining.

“We have essentially four parts that we are contouring on the Haas,” said Wood. “The central part of the heart pump is simply a contour-bladed impeller such as you’d find in any axial pump. Then we have what we call the inflow cage, which is a three-arched bearing support for the motor. We have the outflow stator blades, which are similar to the impeller blades. And then the most complex part is an elbow with both internal and external contouring, which is the outflow end of the pump. That’s the most complex,” Wood emphasized, “it takes us almost six hours to run one altogether.”

The majority of the multi-axis work for the Jarvik 2000 is performed at Transicoil, with many of the prototype and support components being machined at Jarvik Heart in Manhattan. Although the heart pump itself is made of titanium, many different materials are utilized for other components. “We run parts out of titanium, we use a lot of aluminum, we use stainless steels, we use brass and we use acrylics,” Morrow explained.

“The first parts we made on the Haas were part of the implant that would go with the artificial heart, and they were made out of titanium. They were very small pieces, about 3/8" in diameter, and we had to hold a tolerance of two-tenths on the mill work. We’re making injection molds for some of the plastic pieces that will be implanted with the heart; those are out of stainless. We’re also making our own electronic boxes out of 7075 aluminum. We start with a square billet and machine it into a box. Some parts of the boxes – the inside and a lot of the outside – are actually 3-D programming, and we’re using Mastercam to program that, and machining it with a ballnose endmill.

“The Haas handles the materials very well. We didn’t have any problem holding tolerances on the titanium. And the control has been able to handle the large programs that Mastercam generates. We run our tolerances on the Mastercam very tight, so it generates a longer program, but the Haas control handles it very well,” Morrow said.

Transicoil’s Wood agreed, “We all like the control here. We find it very well prompted, easy to follow and very friendly, especially since we were already primarily a Fanuc-type shop.”

About 99% of the programming for the components of the Jarvik 2000 heart is done offline using a combination of standard and custom software. “We typically use a combination of several different software packages,” Jarvik explained. “We use CadKey for the basic CAD layout kinds of things, we use Mastercam as the interface for machining, and we have some custom software that CNC Software, the company that puts out Mastercam, has developed for us especially for machining complex blade shapes.”

According to CNC Software’s John Summers, who has been working with Dr. Jarvik since 1990, the impeller blades of the Jarvik 2000 posed some interesting mathematical challenges. “There’s something peculiar to that part that makes it an unusual machining project,” he said. “The leading edge and the trailing edge are very small radius, and when you’re machining it, it’s important that the part is not turned at all. The cutter goes around the part to give it a cylindrical leading and trailing edge. Then the rest of the blade is similar, because it doesn’t collapse on you – it’s always two points across the blade that are parallel to each other. You could let that blade get as long as you wanted, and it would stay the same thickness and still be wrapped around the hub. We made a 4-axis post processor that was uniquely suited to the part, and made it possible for Dr. Jarvik to machine the blades.”

“Mastercam has been very good,” Jarvik commented, “and it interfaces very well with the Haas machines. We’re using these same programs both at Jarvik Heart and at Transicoil.”

Although the heart is still in the research stage, human trials are expected to begin this year. “We’re doing animal testing now, and we’re getting ready for the first human cases,” Jarvik said. “We’re very confident that, based on the animal data, we can do very well with humans.

“Interestingly,” he continued, “in the history of artificial heart devices, the results in patients have always been better than the results in animals. With animals, there are more infection problems. An animal in a laboratory doesn’t have the kind of medical care and nursing care and all the technology that exists for the patient. And when these devices are applied in a sophisticated, modern hospital, the results are better. So we always expect to do better in humans than we’re able to do in animals. And we’re already able to do very, very well with the animals.”

It’s estimated that 50,000 to 100,000 people each year need an artificial heart or heart-assist device. Even though a small number of devices currently exist and are used as a bridge to transplant, including a variation of the Jarvik 7, they account for only about 2,000 patients per year. “It doesn’t have any major impact on public health,” Jarvik stated. “It’s not until we have a really practical, forgettable permanent device, that it will.”

If the human trials of the Jarvik 2000 heart are successful, the impact on public health will be far-reaching and important. Maybe then the tide will turn in the battle against failing hearts.