Diaphragm Pacing for ALS at a Crossroads Following Conflicting Trial Results
ARTICLE IN BRIEF
Two different trials offer conflicting outcomes data on the use of diaphragm pacing for patients with amyotrophic lateral sclerosis, raising questions about its use for clinical management.
Contradictory results from two trials of diaphragm pacing have researchers debating whether the treatment is beneficial or harmful in patients with amyotrophic lateral sclerosis (ALS) as they struggle to determine how to reconcile the data and counsel patients.
A large nonrandomized trial in the United States found a survival benefit, but in the United Kingdom, patients randomized to receive diaphragm pacing lived only half as long as those randomized to best medical care without diaphragm pacing, leading to an early termination of the trial. The findings from both studies were presented at the Motor Neuron Disease Association Symposium in Orlando in December 2015. Worrying results also brought a randomized trial in France to an early stop, according to an announcement from the investigators.
NON RANDOMIZED TRIAL IN THE US
In the nonrandomized trial, ALS patients were followed as part of the post-approval monitoring program for the diaphragm pacing device, which received a Humanitarian Device exemption from the Food and Drug Administration (FDA) for use in ALS in 2011. Patients were evaluated preoperatively for ventilatory capacity, with 45 percent of forced vital capacity or less as an exclusion criterion.
Preoperative phrenic nerve testing was also done to exclude patients whose innervation of the diaphragm was insufficient to benefit from stimulation. Finally, patients were evaluated during surgery with electrode and video recording to determine whether the diaphragm was responsive to stimulation and whether it was advisable to implant the pacer.
Of the 60 patients evaluated, 54 were implanted with the device, consisting of four electrodes placed at the motor points of the diaphragm and connected to an external pulse generator. Patients had the disease for an average of 636 days. Seventy percent were using noninvasive ventilation, and 15 percent had a feeding tube in place. The surgery was safe, with no serious safety issues except for one pulmonary embolism, according to lead investigator Robert G. Miller, MD, FAAN, an ALS specialist at the California Pacifica Medical Center in San Francisco.
Twenty-four patients died after a median survival of 21 months, which was almost identical to the outcomes reported in the data considered by the FDA for the device exemption, Dr. Miller noted. In order to understand whether diaphragm pacing contributed to survival, Dr. Miller compared all the patients in the study to two separate control groups drawn from recent clinical trials and matched for disease duration, functional disability, and respiratory function. Implanted patients had better outcomes than either set of controls.
Dr. Miller also compared the actual survival of implanted patients to predicted survival based on a model from a large database of ALS patients, called PRO-ACT, for whom survival was accurately predicted based on 40 baseline characteristics. Based on their baseline characteristics, the 15-month survival was predicted to be 40 percent, Dr. Miller said. Their actual 15-month survival was 80 percent (p<1×10−5). Patients who used the stimulator for more than four hours per day did significantly better than those who used it less, he added. “We don't yet know the optimum amount of stimulation with diaphragm pacing.”
RANDOMIZED TRIAL IN THE UK
Results from the randomized trial in the United Kingdom could not have been more different. In that trial, published in the September 2015 issue of Lancet Neurology, 74 patients (of an intended 108) received either implantation of the diaphragm pacing system plus noninvasive ventilation, or noninvasive ventilation alone. Physicians, patients, and caregivers were all aware of which treatment was provided. While there was no preoperative phrenic nerve study, the intraoperative testing procedure was similar to that of the US trial and was performed by experienced physicians with technical consultation from the manufacturer's representatives. Based on the results, all evaluated patients were deemed suitable for implantation.
The procedure itself was safe, said lead investigator Christopher McDermott, MD, an honorary consultant neurologist in the neuroscience department of the University of Sheffield in the UK. There were no deaths within 30 days of the operation, and the eventual causes of death were similar in both groups. But an interim safety analysis showed that the time to death differed, with a median survival of 11.2 months in the implanted group versus 22.5 months in the control group, for an adjusted hazard ratio of 2.27 (p=0.009).
“We conclude that diaphragm pacing is harmful in ALS patients with respiratory failure,” Dr. McDermott said.
A triple-blinded trial in France was also halted early. In that trial, all enrolled patients were implanted before the need for ventilatory assistance, then randomized to begin low-intensity stimulation immediately or to wait. A blinded monitoring committee determined when to begin stimulation in the delayed start group based on the development of respiratory insufficiency. In July, after the announcement that the UK study had been halted, the investigators performed an unplanned safety analysis and found “a statistically significant excess mortality in the group of patients receiving active stimulation,” according to the web site of the Assistance Publique — Hôpitaux de Paris, the sponsor of the study.
Meanwhile, Dr. McDermott and Dr. Miller are cooperating to attempt a pooled analysis of their results to determine whether patient characteristics in the two trials might explain their different outcomes.
“I think the biggest difference in the two populations is going to be that our patients were not as advanced, but why that should then confer harm I find difficult to conceptualize,” Dr. McDermott said.
Data from the French trial and a randomized US trial recently halted before full enrollment, again for safety concerns, may also be included. “I am very happy about the spirit of collaboration in planning this analysis,” Dr. Miller said. “Very clearly, the priority is the patients.”
In the meantime, Dr. McDermott said, “We are not actively pacing anymore in the UK.” Dr. Miller noted that his trial is not enrolling any more patients at the moment, a decision made by the trial's safety monitoring board. “I am telling people I don't think the issue is settled,” he said. “I don't think the British trial is enough to close the door. I continue to be heartened by all the analyses of the US data, and I think we need more information.”
The University of Sheffield study “clearly showed the most dramatic harmful effect we've ever seen in the history of ALS trials,” said Richard S. Bedlack, MD, PhD, FAAN, an associate professor of neurology and director of the Duke University ALS Clinic, who was not involved with any of the studies. “The weight of the evidence is that diaphragm pacing is more likely to be harmful than helpful.”
Given the study findings, he said, “I don't put them in anymore, and I am waiting for publication of the French study to see if we should stop pacing in the few patients who still have them.”
Accurately matching patients to historical controls is very challenging, he noted. “One obvious thing we can't match for is a patient's will to fight. In my experience, patients who wanted the pacers really wanted to fight with every intervention to stay alive for as long as they could,” but without an objective measure of that in the clinical record, it cannot be controlled for in the matching process. The lesson, Dr. Bedlack said, is that nonrandomized trials are important for hypothesis generation, “but you can't take them to the bank.”
Jeffrey Rosenfeld, MD, PhD, FAAN, a professor of neurology and director of the ALS Program at Loma Linda University School of Medicine in California, thinks the jury is still out on diaphragm pacing. “The simple conclusion is that we really don't know if diaphragm pacing is effective, but there is a possibility that a subgroup of patients, when correctly identified, can show us a therapeutic benefit signal. I believe we don't yet know who it is optimal for, but I think we should keep investigating.”
EXPERTS: ON PATIENT OUTCOMES FOR DIAPHRAGM PACING FOR ALS