HF Warning System Surprises by Upping HF Hospitalizations

September 22, 2011

September 22, 2011 (Boston, Massachusetts) — The number of heart-failure hospitalizations in a randomized trial went up significantly for patients with defibrillators featuring a monitor that emits an audible beep when intrathoracic impedance rises far enough to suggest excessive fluid buildup [1]. The sound was to alert patients to impending episodes of decompensation, allowing them to call the doctor and start treatment before they can strike, cutting the risk of hospitalization.

But that didn't happen in the Diagnostic Outcome Trial in Heart Failure (DOT-HF), which investigators reported this week at the Heart Failure Society of America (HFSA) 2011 Scientific Meeting. Compared with outcomes in patients with implants lacking the monitors, patients with them showed up to the outpatient clinic more often (p=0.0001), usually to heed the monitor's audible warning, but their jump in hospitalizations (p=0.022) took the study's organizers off guard.

A potential implication may be that remote monitoring may be preferable to patient alerting.

Also unexpected: when the patients did show up to the clinic, they tended to have fewer signs or symptoms of decompensation than their counterparts without monitors who went to the clinic.

Use of the impedance-monitoring function (OptiVol, Medtronic) and audible-alert system included with the patients' defibrillators "did not improve outcomes. In fact, it increased heart-failure hospitalization in the DOT-HF study, and there are strong indications that the audible alert played a crucial role in that effect," said Dr Dirk J van Veldhuisen (University of Groningen, the Netherlands) when presenting the findings.

"A potential implication may be that remote monitoring may be preferable to patient alerting" for keeping watch over the fluid status of patients with heart failure, he said, although that wasn't studied in DOT-HF.

The trial was published online September 19, 2011 in Circulation [2], with van Veldhuisen as the first author, to coincide with its presentation at the HFSA meeting.

In an earlier, observational study of the OptiVol technology in heart failure, called PARTNERS-HF, rising impedance levels that crossed a predefined threshold were associated with triple the risk of heart-failure events such as decompensation or dyspnea. In that study, presented at the HFSA sessions three years ago and reported then by heartwire , there was no audible signal on which patients could act.

These patients were seen much earlier, before episodes could develop, but often, of course, patients were also seen when it was not really necessary.

van Veldhuisen acknowledged that in DOT-HF, the beeps made by the monitor seemed to drive HF hospitalizations, not prevent them; there was no apparent effect on mortality in patients with the monitors.

"These patients were seen much earlier, before [episodes] could develop," he said, "but often, of course, patients were also seen when it was not really necessary. There was a beep, but it might have been a falsely positive beep."

He suggested that the impedance model as a way to guide therapy in heart failure is still valid, but the audible alert for patients doesn't seem to help.

As the discussant for the DOT-HF presentation, Dr James E Udelson (Tufts Medical Center, Boston, MA) agreed that it may have been the beeps that drove up HF hospitalizations, but there are other possible explanations for the trial's "very negative results."

Negative trials "lead us to question assumptions about pathophysiology and mechanisms," he said. "While the idea that decompensated heart failure is preceded by a prolonged period of subclinical worsening physiologically amenable to intervention is really conceptually attractive, other mechanisms, such as sudden shifts in central blood volume, may often be at play."

The relationship of changes in intrathoracic impedance to clinical heart failure "may be more complicated than we thought," he said. In other trials, the sensitivity of impedance change to decompensation has varied widely. "Those trials suggested many false positives, as well."

Udelson said his remarks were based on his editorial accompanying the DOT-HF journal report that was published at the same time [3].

Pressure-based guidance trials have generally been positive, whereas the impedance trials have not.

DOT-HF had a target enrollment of 2400 patients but halted at 335, the published report notes, "because of slow inclusion and rapidly advancing technology that threatened the clinical value of the tested hypothesis." All those who entered, however, were followed for at least six months.

Entry criteria included heart failure of NYHA class 2 to 4 on optimal medications, LVEF <35%, implantation of a defibrillator or biventricular pacemaker-defibrillator no more than six months previously, and a history of a heart-failure event within 12 months of the implantation.

All the devices were equipped with OptiVol and other diagnostic features; patients were randomized to have all collected diagnostic information available to physicians with the audible alert for the patients or to have the diagnostics and the alert unavailable to physicians and patients. The 335 patients available for the intention-to-treat analysis were followed for an average of 15 months.

For the 168 patients in the OptiVol group compared with the 167 in the control group, the hazard ratio (HR) for the primary end point of all-cause mortality or HF hospitalization was increased but fell short of significance at 1.52 (95% CI 0.97–2.37, p=0.063). The HR for HF hospitalizations alone was significant at 1.79 (95% CI 1.08–2.95, p=0.022) and for death from any cause was nonsignificant at 1.24 (95% CI 0.63–2.44, p=0.54).

Patients with OptiVol had three times as many unscheduled outpatient clinic visits as those without the monitor, yet they less often had symptoms or signs characteristic of decompensation.

Disconnect Between Frequency of Outpatient HF Clinic Visits and Visits Involving Actual Decompensation Symptoms or Signs, Optivol vs No Optivol

Parameter OptiVol (n=168) Control (n=167)
Outpatient clinic visits (n) 250* 84
Patients with actual symptoms/signs of decompensation (%) 16 26

*p=0.0001 vs control

It's conceivable, Udelson said, "that impedance is the wrong marker, that maybe monitoring pressures is better." He pointed to the CHAMPION trial, in which a pulmonary-artery pressure sensor implanted in patients with NYHA class 3 heart failure and a history of decompensation cut hospitalizations by 30% compared with standard care, and to the COMPASS-HF trial, when a different pressure monitor was associated with a 22% reduction in heart-failure events overall and 41% for patients in NYHA class 3.

"So, pressure-based guidance trials have generally been positive, whereas the impedance trials have not." In both of the pressure-monitor trials, he noted, patients were blinded to pressure changes; that information went directly to the doctors.

"It may be that, despite the fact that in contemporary practice, we very highly value shared decision-making with our patients and involving them in decisions about their care," Udelson speculated, "in this situation, involving patients in the information-treatment loop may be an incorrect strategy."

DOT-HF was sponsored by Medtronic. van Veldhuisen discloses receiving research funding and/or consultancy fees from Medtronic and Biotronik. Udelson discloses being a steering committee member for an unrelated Medtronic-funded study.