Thursday, July 2, 2009
Palliative sedation and shortened survival
Annals of Oncology has a paper trying to look at whether palliative sedation is associated with shorter survival. It's a prospective, multi-center, single-country (Italy) study which prospectively matched ~270 cases (cancer patients admitted to hospice units who received palliative sedation) with similar 'controls' (similar cancer patients admitted to the same hospice units during the same period). Controls were matched for gender, age class (less than 65 and greater than 65 years), reason for admission (psychosocial, uncontrolled symptoms, and terminal phase), and Karnofsky performance status (KPS) subdivided into three classes (10–20, 30–40, and 50 or more). Patients who received PS did so based on standard practice by their treating physicians; there was also no standardization of PS treatments.
As you'd expect by the design patients were well matched; even though not deliberately matched the two groups ended up having similar Palliative Prognostic Scores, and so should have had grossly similar survival.
They admit that there are a variety of practices which fall under the umbrella of 'PS' and as far as I can tell they included anyone in PS group who received anything which the treating physician labeled as 'PS.' Along those lines, only 25% are described as receiving 'continuous, deep' sedation (ie what historically was described as 'terminal sedation'). There is a confusing array of data presented (some patients receiving mild sedation, others intermittent, some continuous but not deep, a whole variety of meds involved including opioids, etc.).
About 80% of patients received PS for delirium/agitation, with the rest being dyspnea and pain (and 6% for psychological distress only). The survival curves of both groups were identical (median survival in the 10 day range). They did not present separate data for just the 'continuous/deep' sedation group (and their controls). Cessation of artificial nutrition or hydration was not measured.
Their conclusion (from the abstract): "PST does not shorten life when used to relieve refractory symptoms and does not need the doctrine of double effect to justify its use from an ethical point of view."
Some thoughts here.
1. While I may think their conclusion is correct, that's not because this study particularly supports it. Most of the debate about PS has been about 'continuous/deep' sedation/'terminal sedation': essentially deliberately inducing a pharmacologic coma (state of unresponsiveness), with the intention of maintaining that coma until death (usually, although not necessarily, accompanied by cessation of other life-sustaining medical treatments including artificial nutrition and hydration). Only a quarter of these patients, apparently, had this flavor of PS, and we aren't presented with a separate analysis of these patients. Without doing that this is like arguing that light, intermittent sedation doesn't hasten death but that's not really a burning clinical or ethical debate currently.
2. This is a group of patients with a median survival of ~10 days. Even though this is a relatively large study, unless any intervention had a major impact on mortality demonstrating a difference in survival is nearly impossible. If they had given all the PS patients 1gm of pentobarbitol and a bucket of succinylcholine and found that median survival in the 'active' group was 5 minutes vs. 10 days for the controls that'd be one thing but this is not. The real point here is imminently dying patients are imminently dying patients and we're never, ever going to show a mortality difference between PS and non-PS groups of those patients. This sort of study design would be the way to do it - a randomized controlled study trial could not happen for ethical reasons. You could fancy this study up with a more elaborate propensity score matching process, etc. but it'd be essentially be the same study design. It's not going to show any difference however.
3. Is there a real, significant ethical question to be answered about continuous/deep sedation given for intractable symptoms in otherwise imminently dying patients? And is it in fact an important question to even be asking/worrying about - the potential for shortened life? Assuming CDS-PS is used as a last resort (ie nothing else has worked adequately, leaving aside questions about what that might actually mean) it seems the available therapeutic scenarios we have are:
- Staying the course: patient dies with uncontrolled suffering in ~10 days.
- CDS-PS without any life-prolonging medical care: patient dies peacefully in 10 days and yes, perhaps it's possible they would have lived 11 days without CDS-PS, or hell maybe we could have made them even live longer like a 12 days or 3 weeks if we mechanically ventilated them, put them on pressors, threw in some CVVH, and even tried a little ECMO.
- Actively hastening death: applying novel medical interventions with the intention of ending life/shortening life; patient dies in less than 10 days with, well, no more suffering subsequently.
I should not be naive: there are of course plenty of people for whom the thought of potentially 'hastened death' (really it's 'not-prolonged death') via deliberating sedating patients or using meds with sedating side effects is troubling. Research like this can be reassuring which is a good thing. I'm arguing however is that what is 'really' needed is not reassurance that these practices probably don't influence survival but engaging with clinicians who have these concerns about the fundamental issue: that it's ok not to prolong dying, that a peaceful, comfortable death should be the preeminent treatment goal when someone is dying and we don't need to twist ourselves into ethical knots about it.
4. The real ethical question about CDS-PS is whether it is an appropriate option for patients who aren't imminently dying (and I'm not going to define exactly what I mean by that), including for patients who have intractable psychological and existential suffering. In these cases, CDS-PS is almost certainly life-shortening and the question is is that acceptable? How much 'life-shortening?' How much suffering? This is a question that can't be answered by research.
Hat tip to Lyle for his thoughtful comments about this post.