Friday, June 15, 2007
Neurology has an important article looking at prognosis in intracerebral hemmorrhage which finds that early treatment limitations are independently associated with mortality (related editorial here. ) The data come from a large, regional study of ICH outcomes in Texas, and the researchers reviewed charts looking for 'care'-limitations (DNR orders, decisions to not escalate care, withdrawal of life-prolonging treatments) in non-brain dead patients. Long-term mortality was also followed - median length of follow-up was 417 days. They looked at 270 ICH events total: of those, 93 had early (<24h of admission) treatment limitations vs. 177 who did not. Those with early TL's were older (77 vs 70 years), sicker (lower Glasgow Coma Score, larger ICH volume, more likely to come from a NH), and had much higher in-hospital mortality (75% vs 25%). Long-term survival was also very different (~65% survival at 3 years without TL vs 20% survival at 3 years with TL); all the difference in survival however occurred immediately (survival curves are identical - nearly flat - once past the initial event). None of this is surprising - the patients with TL were much sicker. What is notable is that after controlling for age, gender, ethnicity, initial coma score, ICH volume, extent and site of hemorrhage, early treatment limitations were still significantly associated with in-hospital and short & long-term mortality. Implication: the TL patients were in fact considerably sicker (but in a way that wasn't controlled for in the multivariate analysis...for instance medical comorbidity wasn't accounted for at all) or many of the TL patients would have lived much longer without those treatment limitations. Probably both of these are true; withholding or withdrawal of life-prolonging care is just that. The authors conclude: [E]arly predictions of poor prognosis and subsequent reductions in care could result in withholding care from individuals with the potential for meaningful recovery. Physicians may therefore wish to exercise caution prior to recommending limitations in the aggressiveness of care in the initial hours after ICH presentation.
Translation: don't withdraw/limit life-prolonging care immediately because you may be wrong. Those recommendations, then, are the real challenge of the study. What is missing from them though is that while TL may lead to earlier death that may very well be consistent with the patient's values - not have their life prolonged after such an insult if their quality of life afterwards was not going to be acceptable to them. Predicting that is, of course, difficult and weighing out the chances of that happening with a patient's desire not to have potentially burdensome care is the real trick and this research is not designed to speak to that. They mention 'potential for meaningful recovery' but this wasn't measured here - just survival - although if you're dead you don't of course have that 'potential.' To me, it seems like survival is the wrong outcome to measure in these situations. It's an easy outcome to measure to be sure, but what most patients want to know is their likelihood of being restored to health (to an acceptable level of health and function) and not their chances of solely being kept alive. My thinking that that should be the relevant outcome doesn't make it possible to define or measure well or 'operationalize' but one can still think it.
There are some patients who would have recovered well but don't due to treatment limitations, just as there are some who would have not wanted their lives prolonged who nevertheless received full life-prolonging care. Without perfect knowledge (or even much better knowledge than we have now) these 'mistakes' will be made. But are these 'mistakes'? Is that even the right way of thinking about it? Assuming we will never have perfect knowledge, if we (medical professionals) honestly and with humility talk with patients/families about both the limits of our knowledge to predict and the limits of our ability to heal, earnestly try to understand a patient's values and goals and limits, and then make the best decision possible - that is good medical care and not a 'mistake.' Holding ourselves to perfectly informed decisions is a positivist folly.
So, do treatment limitations in ICH increase your chance of dying? Yes - but that's not the right question.
The British Journal of Cancer has an article comparing buccal fentanyl with IV morphine in a prospective, randomized, non-blinded trial for breakthrough pain in cancer. Despite involving only 25 patients (inpatients in a palliative unit) this was an astonishing complicated trial & results and I'm not quite sure what to make of them. I'm more mentioning this for those of you with a particular interest in transmucosal fentanyl. The major question it is addressed is whether it's safe to start buccal fentanyl in a patient using a dose which is a percentage of their overall opioid dose (many of us choose a starting breakthrough dose of ~10% of the daily opioid dose) as opposed to using the lowest buccal fentanyl dose and titrating up (this latter practice has been proposed because in other trials of buccal fentanyl the effective breakthrough dose seemed to have no relation to the prior opioid dose). In this study they chose 200mcg of fentanyl per 60mg of oral morphine equivalents daily dose as their breakthrough dose and it seemed to work and be safe and effective. That is, for someone on 180mg of oral morphine equivalents daily their data suggest 600mcg is a safe breakthrough dose of buccal fentanyl.
And finally, courtesy of alert reader Dr. Steve Rommelfanger, The Onion once again has an offensive and funny 'end of life' related piece: " I've got some bad news, and I've got some hilarious news." Also note this fake headline from 2002 .