Friday, September 18, 2009
American Journal of Respiratory and Critical Care Medicine has a paper about the prognostic importance of delirium in ICU patients, and continues the thread of research we've been following recently about delirium and prognosis.
This study was a single-center (Yale) prospective cohort study of older (over 60 years) adults in an ICU, in which delirium was prospectively and rigorously assessed (by research nurses, using the CAM-ICU scale); ~300 consecutive eligible admissions were enrolled. They note that they included patients with 'coma/stupor' as delirious; I initiatlly assumed this implied patients who were spontaneously stuporous/comatose (due to a high grade encephalopathy from sepsis, etc.) as opposed to pharmacologically induced sedation/stupor. However as far as I could tell there were no such exclusions and these data apply to patients with delirium/coma from any cause. They basically looked at the number of days a patient was delirious/altered in the ICU, and correlated that with survival.
Patients (mean age 74 years, 31% with a diagnosis of dementia, 54% intubated, 82% received opioids or benzos in the ICU at some point, median ICU LOS 6 days, meidan hospital LOS 11 days, and median days of ICU delirium was 3). 16% of patients died in the ICU and 50% died by 1 year.
Yes, that's 50% 1 year survival for older patients admitted to an ICU, with or without delirium.
They did a mutlivariable Cox regression analysis and found that number of ICU delirium days were independently associated with shorter survival (hazard ratio); age, comorbidities, and severity of ICU illness were also independently associated with worse survival. Eye-balling median and 1-year survivals based on days of ICU delirium (this is me eye-balling a Kaplan Meier curve): median survival for ICU delirium lasting 5 days or longer was about 2 months with 1 year survival being about 35% for those with delirium 5+ days and 45% for those with 3-4 days. Patients without delirium did much better: 1 year survival was 75% (as far as I can tell the number of patients this represents is not shared).
This is further supporting evidence to the now vast pile of data indicating that delirium in multiple settings is a poor prognostic marker. To take these data at face value, being older, in an ICU, and having alterations in your mental status for more than 4 days implies a 50% 2 month mortality and 2/3 1 year mortality (and any ICU stay implies only a 50% chance of surviving a year). It's important to acknowledge that despite the good quality of the data collection for this type of research - prospective, used a validated assessment tool, all-comers were offered enrollment, etc. - it is still from a single institution and the absolute value of these numbers is not easy to generalize. That said, whether it's 2 months or 4 months or whatever: it's bad, and this study is as good as any recently to show that.
What do we do with this? I have always hummed and hahed about how to apply this sort of research at the bedside. We already know if you're 75 years old, and in an ICU, for any reason - it's a marker of high 1 year mortality. Which is not to say that you're likely to die in the ICU, and that going to an ICU isn't going to 'save you.' Most of these people survived the ICU, although most who had rocky ICU stays died within a few months. That, to me, is the real story - we can save you, but often for a relatively short period of time (and who knows what quality of life). Patients want to live, and be 'saved,' and all that good stuff, but we are doing them a disservice if we don't let them know that while we can 'save them,' and all that good stuff, it's usually not going to be for a long and healthy life. It's time to begin planning, and sorting through how you want to spend that time. A wise mentor once told me when talking about The ICU and older patients is that that the big question is not can this patient survive the ICU (most do and we have incredibly sophisticated systems in place to get people out of ICUs alive) but that can they survive it for what? And for how long? And while most patients want to live, knowing that they might be facing going through that again, and with likely little long-term benefit, many will choose not to do it again.
Anyway - the big etiologic question here, which may have a long term impact on how patients are treated, especially as pain and symptom meds do, indeed, cause/worsen/lengthen delirium, is whether the delirium itself causes the worse prognosis, or whether it's a passive marker of poor prognosis. If it's the delirium itself then anything to prevent or shorten it will make people live longer presumably: that will 1) hopefully be an impetus to actually do large/good trials of delirium treatment (in ICU and anywhere else - we need 'em), 2) raise the spectre of withholding symptom meds in hospitalized patients as they could be perceived as causing/worsening/lenghtening delirium and thus worsening long term prognosis?
(Thanks to Dr. Bob Arnold for alerting me to this.)