Tuesday, September 20, 2011

Thoracic Impedance Monitoring Fails to Limit Heart Failure Admissions

Imagine being able to detect heart failure before it happens. The potential to offer significant cost reductions in health care might be significant.

So device makers created a way that implantable defibrillators could measure a drop in impedance between the pulse generator can and the lead in the patient's heart. Lower impedance should mean more water in the chest, suggesting fluid overload.

It seemed, well, so logical.

But with today's release of the DOT-HF trial, it seems that all that thoracic impedance information contained in an ICD, when used to alert the patient to a fluid accumulation in their chest, has the potential to create more rather than fewer hospitalizations:
METHODS AND RESULTS: We studied 335 patients with chronic heart failure who had undergone implantation of an implantable cardioverter-defibrillator alone (18%) or with cardiac resynchronization therapy (82%). All devices featured a monitoring tool to track changes in intrathoracic impedance (OptiVol) and other diagnostic parameters. Patients were randomized to have information available to physicians and patients as an audible alert in case of preset threshold crossings (access arm) or not (control arm). The primary end point was a composite of all-cause mortality and heart failure hospitalizations. During 14.9+/-5.4 months, this occurred in 48 patients (29%) in the access arm and in 33 patients (20%) in the control arm (P=0.063; hazard ratio, 1.52; 95% confidence interval, 0.97–2.37). This was due mainly to more heart failure hospitalizations (hazard ratio, 1.79; 95% confidence interval, 1.08–2.95; P < 0.022), whereas the number of deaths was comparable (19 versus 15; P = 0.54). The number of outpatient visits was higher in the access arm (250 versus 84; P < 0.0001), with relatively more signs of heart failure among control patients during outpatient visits. Although the trial was terminated as a result of slow enrollment, a post hoc futility analysis indicated that a positive result would have been unlikely. CONCLUSION: Use of an implantable diagnostic tool to measure intrathoracic impedance with an audible patient alert did not improve outcome and increased heart failure hospitalizations and outpatient visits in heart failure patients.
Postulating why this unexpected result was seen, Dr. James E. Udelson in an accompanying editorial suggested several reasons:
In the DOT-HF trial, among the episodes of HF hospitalization, only 60% were preceded by an alert condition, where impedance readings fell below the prespecified threshold indicting risk. In other words, the sensitivity of the impedance alert to predict a HF episode was only modest. This is actually consistent with previous studies, and false-positives seem common, as well.
Almost half of the outpatient visits in the treatment group were a result solely of an alert, which may have resulted in a drive to more intervention and possibly hospitalization.

Moral of the story: more data from ICDs is not necessarily better when it comes to predicting heart failure exacerbations.

-Wes

References:

van Veldhuisen DJ, et al. "Intrathoracic Impedance Monitoring, Audible Patient Alerts, and Outcome in Patients With Heart Failure." Circulation 2011 http://circ.ahajournals.org/content/suppl/2011/09/18/CIRCULATIONAHA.111.043042.DC1.html

Udelson JE. "T.M.I. (Too Much Information)?" Circulation http://circ.ahajournals.org/content/early/2011/09/19/CIRCULATIONAHA.111.067819.citation

2 comments:

Tim Hulsey, MD said...

That's one way not to make a light bulb!! Keep trying.

TakeItTo11 said...

Numerous studies have shown that thoracic impedance alone is not the best way to use the tool. This study only validates what we already know. Use OptiVol with other indicators such as HR variability, Night HR, etc. See Perego, Landolina, Vergara work in Journl of Interventional Card Electrophysiology. Dec 23 2008. "Implantable CRT device diagnostics identify pts with increased risk for HF hospitalizations"
C'mon guys, you should remember this!