Monday, August 27, 2007
3 on heart failure...
Journal of the American College of Cardiology has a trial looking at left-ventricular assist devices (LVADs) as destination therapy in patients with advanced heart failure. LVADs, in the simplest terms, are implantable devices which augment the pumping of a failing heart; they are relatively well accepted as a temporizing measure while people are awaiting heart transplants. They have also been controversially proposed for people who are ineligible for heart transplantation - as 'destination therapy.' The current study is a non-randomized, observational study of this. The patients (37 in the LVAD group, 18 in the medical management group) had Class IV heart failure symptoms (rest symptoms), severe systolic dysfunction, and had 'failed' inotrope weaning at least twice. They were then offered LVADs - most received them; those that didn't either declined them, had 'inadequate identifiable financial resources' to receive one, or had a mechanical aortic valve replacement (which was identified as a contraindication to LVAD placement). So, yes, the control group had different inclusion/exclusion criteria than the treatment group.
As you'd expect from an uncontrolled trial the control group was different than the treatment group. They were sicker: higher BUN and lower serum sodium levels (significant) and trends toward worse creatinine and wedge pressures. The LVAD group's survival was poor but substantially better than the control group: 6 month was 46% vs 22% and 12 month was 27% vs 11%. The authors mention quality of life was better in the LVAD group but don't present that data or baseline quality of life information.
The authors not only conclude that LVADs look a lot better than medical management for this patient population but suggest it is now no longer ethical to do a randomized study because one cannot reasonably argue there might be equipoise with medical management. The first statement seems reasonable: based on these data LVADs may prolong life (and we are reassured improve quality of life although those data weren't deemed important enough to actually present). While there's reason to doubt this; most notably due to the fact that the control group patients were sicker, had different inclusion/exclusion criteria, and it's probable that there's significant confounding in the decision to receive an LVAD or not (patient 'choice' and not having financial resources...when was the last time you saw that in a treatment trial as a reason not to receive the therapy?), the magnitude of the effect is pronounced (survival was doubled). But to say that these data, with all their associated problems (& I didn't even mention the small numbers), are sufficient to forever exclude an actual controlled trial on ethical grounds (lack of equipoise) seems a stretch.
My sniping aside, we are almost certain to see more of LVADs in the future, and they will further complicate the already complicated decision-making in heart failure, and should be a challenge we are ready for. And if there's one thing this study does clearly demonstrate it's the extraordinarily high mortality of these most-advanced heart failure patients: a prime patient population for palliative care priorities. To this end, I'd love to see the quality of life data...assuming they're legit (particularly that they suggest a clinically and not just statistically important improvement in QOL) this will highlight that for heart failure (perhaps more than many other terminal illnesses...?) distinguishing between life-prolonging interventions and quality of life interventions is often impossible. They're the same thing, insofar as getting the heart to pump better generally makes people feel better! The implications of this for hospice care and hospice's financial model will be interesting.
(Related editorial here.)
Full-sized LVAD image (& further discussion) at WebMD, or click on the image above.
Annals of Internal Medicine has systematic review of implantable cardioverter defibrillators for people with systolic dysfunction, which concludes they reduce mortality in patients with class II & III symptoms but there are inconclusive data for the healthiest and sickest patients (classes I & IV). Absolute risk reductions and numbers-needed-to-treat are absent from the analysis.
The authors also noted these concerns, despite the mortality benefits of ICDs, and despite my argument in #1 above:
"However, our data on patient and device-related complication rates highlight the perhaps underappreciated risks of ICDs, particularly in light of 3 findings. First, three quarters to two thirds of ICD recipients in the observational studies received no therapeutic ICD discharges, and only 5% to 12% of trial participants received an appropriate shock per year. Second, the frequency of inappropriate shocks was surprisingly high, and at least 1 study has demonstrated that inappropriate shocks are associated with an increase in risk for death (hazard ratio, 1.97 [CI, 1.29 to 3.01] in the Sudden Cardiac Death in Heart Failure Trial). Third, although the studies we reviewed infrequently report quality-of-life outcomes, some studies have shown that quality of life declines in many ICD recipients, especially those who experience frequent ICD firings. Not unexpectedly, patient anxiety and psychological distress scores increase substantially after an ICD shock or after publicity about device recalls."
(Annals also has an analysis of how quickly systematic reviews become outdated for those of you who are interested. Answer: pretty quickly.)
Heart has a qualitative study looking at palliative care needs of heart failure patients, their caregivers, and their health care providers. It's based on focus group-type interviews. The themes you'd expect emerge did emerge: patients not conceiving as their disease as terminal; cardiologists never 'running out' of disease-modifying interventions to try, and not being comfortable talking about mortality, particularly in light on the difficulties prognosticating for an individual patient; and so discussions about end of life concerns & care rarely occurring. What perplexes me about this is that while it's near impossible to prognosticate with any precision for most patients - more than "likely to live many years" or "not likely to live many years" - for those patients who are not likely to live many years what is likely to happen to them in that period of time is generally clear: ups and downs but with overall progressive functional decline, punctuated by episodes of critical illness, with each episode carrying high risk for death. Preparations for crisis and discussions of goals & treatment limitations in light of functional decline seem very pertinent, despite not being able to put a 'number' on the prognosis.