A Reliable Test to Detect Impending Pump Failure during Long-Term Support on the Novacor® N100 Left Ventricular Assist System

Abstract

The Durastudy is a new surveillance protocol to detect impending pump failure in Novacor® N100 Left Ventricular Assist Systems implanted for the long term. Our patient, a 54-year-old man with a history of dilated cardiomyopathy and contraindications for heart transplantation, received a Novacor pump in May 1995 and did not experience sustainable ventricular recovery during the subsequent 3 years. After more than 3 years of support, symptoms of pump wear were detected in this patient, through application of the Durastudy protocol. This enabled us to electively exchange the pump at 3.8 years. Our patient remained in New York Heart Association functional class I until he died in July 1999 of causes unrelated to the pump, after some 1,514 days of support. This, we believe, still constitutes a world record.

As of April 2001, 1,233 patients had received Novacor® N100 Left Ventricular Assist System^* (LVAS) implantations. Although applications in the United States are restricted to “bridge to transplant,” the system is not restricted in Europe and 4% of applications have been as a “bridge to recovery,” with a further 3% implanted as definitive therapy. Eighty-eight patients have been supported for more than 1 year, and 20 of these have been supported for more than 2 years. Since 1994, with the introduction of the wearable system, increasing numbers of patients have been discharged home, for longer periods of time, so that in 2000, 80% of patients spent more than 85% of their support time outside the hospital. This experience prompted the manufacturer to initiate a project aimed at developing a simple, reliable, and noninvasive method of surveillance aimed at detecting impending pump wear-out.

In vitro “life-testing” of the Novacor LVAS had revealed a single failure mode of wear-out, which involved failure of the main bearings at 4.4 ± 0.7 years, under nominal test conditions. ¹ Bearing failure, related to reduced lubrication, was accompanied by a number of characteristic signatures in electronic tuning and power parameters to the device, which were manifest up to 6 months before functional failure. The Durastudy protocol ² was therefore based on these laboratory findings and has been evaluated, since 1997, by carrying out 60 tests in 40 patients from 11 European centers. The protocol involves connection of the patient to a Novacor monitor, diagnostic manipulations using the monitor, a simple test during exercise to monitor flow dynamics across the valves, and examination and sampling of any deposits in the driveline. This procedure takes about 20 minutes and has been well tolerated by all patients. Durastudy evaluations were carried out on most patients whose LVAS devices had exceeded more than 1 year of implantation, and they were repeated at approximate 6-month intervals thereafter, unless a given LVAS showed a need for more frequent monitoring.

This paper describes the course of 1 such chronically managed patient, in whom impending failure of the LVAS was detected through application of the Durastudy protocol.

Case Report

A 54-year-old man, with a history of dilated cardiomyopathy and contraindications for heart transplantation, underwent implantation in May 1995 of a Novacor N100 LVAS. After postoperative recovery, the patient returned home (200 kilometers from the hospital) and led a near-normal life in the community. ³ In June 1998, he experienced a series of febrile episodes that were eventually traced to a Staphylococcus aureus infection of the inflow and outflow valve conduits. ⁴ These conduits are modular on the Novacor, which facilitates replacement. After replacement, the infection was eradicated and the patient had no additional problem.

In this patient, we began the Durastudy protocol in April 1997 and carried out repeat tests in January, May, August, and December of 1998 (Table I). The analysis of results showed evidence of increasing wear, first detected in May 1998, approximately 3 years after implantation.

Table I. Results from Sequential Durastudy Measurements

A Durastudy performed in mid-December 1998 at 3.56 years (corrected to a pump rate of 100 bpm, this prorates to 4.06 years) revealed corrected tuning changes of −33 units (corresponding to 120 days of remaining life), power changes of −1.36 amps (corresponding to 200 to 400 days of remaining life) and vent-tube deposits that consisted predominantly of iron oxides (which suggested only main-bearing involvement). Together, these findings suggested that the remaining period of reliable function was between 3 and 5 months. After careful consideration, we decided to carry out an elective pump exchange, before overt wear-out symptoms were experienced.

Reoperation was undertaken on 27 January 1999, on the 1,342nd day of implantation (3.8 years). A conventional median sternotomy was required. The patient was placed on supportive cardiopulmonary bypass (CPB) via aortic arterial cannulation and single 2-stage venous cannulation. An incision, approximately 15 cm long in the 11th intercostal space, provided access to the inflow graft and left lateral aspect of the pump. Once CPB support was established, both the inflow and outflow grafts were clamped close to the valved conduits. The connectors were undone using standard Novacor wrenches, the driveline was transected and capped, and the old pump was removed. The driveline for the replacement pump was tunneled along a tract approximately 2 cm cranial to the original driveline tract. The pump was inserted into the pocket and connected to the inflow graft. The inflow clamp was partially released to facilitate priming, and then the outflow grafts were connected. De-airing was accomplished by means of a needle vent placed in the outflow graft and the use of single strokes of the pump, until no more air was evident. The patient was weaned from CPB onto full pump output, in the usual way. The remnant of the original driveline was mobilized and removed, and the incisions were closed after we placed 2 drains in the pump pocket, and the usual thoracic drains. The patient returned to the intensive care unit, stable with good hemodynamics and a pump output of approximately 6.0 L/min. The patient slowly recovered and, after 1 month, was discharged home. Our patient remained in New York Heart Association functional class I until he died in July 1999 of causes unrelated to the pump, after some 1,514 days of support.

Discussion

The growing supply-and-demand disparity in regard to donor organs has made it necessary to look for alternative therapies for the treatment of patients with terminal heart failure. These comprise high-risk conventional surgery, ⁵ partial left ventriculectomy, ⁶ and chronic implantation of left ventricular assist systems. ³

Available LVAS devices are still sub-optimal for true chronic support. However, the Novacor LVAS appears to offer a durable and reliable solution, and provides an acceptable quality of life for patients with no other option. ⁷ The advent of the Durastudy surveillance regimen has added another dimension to the management of patients with chronic LVAS implantations. This approach enabled us to monitor our patient with confidence and to schedule replacement electively, when the pump approached the end of its life.

The electronic tuning and power changes occur because the opposing faces of the solenoid drive motor fail to remain parallel, as the bearing clearances increase with wear. This results in a reduction in the effective stroke volume and in a subsequent reduction in tuning values and power requirements. A reduction of tuning values of >10 units and a decrease in power requirements are associated with approximately 6 months of remaining life. As wear progresses, the changes accelerate: when tuning values are >30 units and are accompanied by a rapid rise in power requirements (resulting from the tendency of the solenoid to “bounce out”) and by the appearance of iron residue in the driveline filter, it is likely that less than 3 months of reliable life remain in the pump.

By extrapolation from the manufacturer's in vitro data, ¹ our patient's pump had operated for the equivalent of 4.06 years at the time that we made our decision to replace it. This is a little less than the mean duration of nonstop operation of the 12 Novacor in-vitro test pumps, which failed at a mean duration of 4.4 ± 0.7 years (minimum, 3.2 years; maximum, 5.6 years). However, we were unwilling to have our patient exposed to the terminal behavior of pump wear-out, which is characterized by increasing misfires and alarms, so we electively replaced the pump before this stage.

Analysis of our patient's explanted pump at Novacor was consistent with the bearing wear (Fig. 1) that the company had anticipated, and the débris typically associated with wear of the bearing and shaft was distributed throughout the motor section. There was increased play between the shaft and bearings to a degree that hard contact between the solenoid poles of the armature probably would have occurred within a few more weeks of operation. This confirmed that our decision to replace the pump was appropriate and that the Durastudy findings were prognostically accurate.

Fig. 1 A new main bearing (left), in comparison with the main bearing from our patient's device after 3.8 years in clinical operation. Note the lack of lubricant and the uneven rollers.

To the best of our knowledge, our patient's 1,514-day survival on a left ventricular assist system constitutes a world record and lends credence to the use of such devices as definitive therapy when other options have been exhausted.

Acknowledgment

The authors would like to thank World Heart Inc. for reviewing this manuscript to ensure the accuracy of its technical detail.