Monday, November 16, 2009
This will be the last of the Mag Citrate posts for now. And after reading this monster you'll thank me. That said, these are all notable papers, each which I hope will be of interest to at least some of our readers: the progno-wonks, the teaching-filists, the EBM-purists, and the Terrible Things We Do To Our Patients-people.
(Images chosen because they are pretty and available under a CC license. See here for an explanation of the photographic technique used.)
Cancer has a paper about prognosis in advanced biliary cancers. The data come from ~200 patients with unresectable biliary cancers (gallbladder, intra & extrahepatic cholangiocarcinoma) who were enrolled in a variety of studies earlier this decade (it's a Korean study group; unclear if all the patients were Korean). Data were collected both prospectively and retrospectively. All of the studies involved chemotherapy (ie these were patients 'well enough' to be enrolled in chemotherapy studies, which means that if anything these findings overestimate survival). Mean age was 59 years; most patients had metastatic disease and a quarter of included patients had recurrent/metastatic disease (meaning they initially underwent resection, and 'enrolled' in this study at the time recurrence/metastasis was diagnosed).
Gross survival data were presented: median survial was 7 months, 6-month survial 57%, 1-year survial was 28%, 2-year ~10%, and 3-year and beyond was less than 5%. They did a multivariate regression analysis, and in the end having intrahepatic cholangiocarcinoma (vs the other biliary tract cancers), liver metastases, poor performance status, and metastatic disease (vs. locally advanced) were all associated with worse survival. The authors tied this up into a prognostic index to stratify 'better' and 'worse' prognostic groups. Frankly, though, it's all dismal, and without further investigation of this prognostic index, these gross survival data are the most helpful.
Thorax has a review on the role of opioids for dyspnea in COPD. It's a narrative review, which gives some history, reviews the literature, and discusses barriers to opioid use for dyspnea. One major point they make is that despite the consistent research findings that opioids are effective for dyspnea in COPD, long-term use has not been studied at all (most trials involved single doses or were for less than a week). The authors discuss their approach to using opioids for dyspnea - one could describe it as starting extremely low (e.g. 1-2.5 mg of oral morphine) and going extremely slowly; quite a different approach than how most of us use morphine for pain.
They also outline a wide-ranging research agenda. In it they do mention the issue of patients with pre-existing hypercapnea (CO2 retention). I hope someone addresses this, as it remains an important question at least in my practice. The aforementioned research, all of which supported opioids' safety/efficacy (including that they don't seem to induce hypercapnea), was all done in patients without chronic hypercapnea, and I think it remains a real question if opioids are safe in this population (assuming patients have a longer prognosis and goals which aren't strictly symptom-alleviation). (See here for a good systematic review of the research ([it misses a few recent studies, none of which however answer the above questions]). If you've got a patient with a baseline pCO2 of 60, an indeterminate prognosis, chronic and distressing dyspnea which hasn't responded to other interventions - do you give them opioids?
On the 'real evidence' front, NEJM has a sobering analysis of selective reporting of outcomes regarding the use of gabapentin for off-label purposes (most of these were investigating gabapentin's use for chronic and neuropathic pain). They note that much of the information in this analysis comes from documents which were made available as part of an investigation into the manufacturers' off-label marketing of the drug. Internal research protocols were evaluated, and, among other things, compared to what was actually published/made public from the studies. They found that over half of the publications of these trials (8 of 12 published trials) either published secondary trial outcomes as primary outcomes (ie in the research protocol an outcome was designated as a secondary one, but published as if it was the primary outcome) or published secondary outcomes in a manner indistinguishable from the protocol designated primary outcome, or published novel secondary outcomes (ones that weren't ever specified in the research protocol). They also demonstrate that, not surprisingly, for most of these publications which fudged the primary outcome it made gabapentin appear more effective.
Their bottom line: We are concerned that the reporting practices observed in our analysis do not meet the ethical standards for clinical research or maintain the integrity of scientific knowledge. Fair and honest treatment of patients enrolled in clinical trials of any kind requires full, open, and unbiased reporting. Journal publication, a formalized platform for scientific discourse and dissemination of knowledge, should not be used as a marketing tool for off-label drug use.
It's impossible to know how widespread these practices are, and they point out that all this comes from the fact that gabapentin's manufacturers were investigated for inappropriate off-label marketing of the drug (ie we have reason to believe already the manufacturers weren't playing by the book). Nonetheless it's a real warning that the grain-of-salt that we all take with purely industry funded research is not just paranoia or nit-picking skepticism. Of real importance is the reality that there remains a good amount of either unpublished or 'grey' published research - and that what we see tends to be the Drug X Works publications, particularly if research into the drug is industry-only supported. I think gabapentin works, probably most of us who read this think so, albeit modestly so - but what we know about how well it works is probably overstated, however.
JAMA recently published a report looking at broad demographic trends in hip fractures in older patients (over 65 years), which has some prognostic data (we visited this topic a couple years back: see here, item 3). Huge study, using 20 years' worth of Medicare claims data. The gross survival data were in the ball park of the study I just linked to: 1 year survival was ~22% for women, 32% for men. So not quite the sentinel event for older patients with advanced dementia (see here), but not great either.
Archives of Internal Medicine recently published a brief report of a survey of physicians' experience with and satisfaction with chaplains. The data come from a large national survey of physicians (N=1100, all specialties and practice types, about a 62% response rate). Basically most physicians had worked with chaplains and were satisfied with the work chaplains did. A couple things I found interesting about this. First was its use of 'R/S' throughout to denote 'religion/spirituality.' Something about this bugged me, which is weird since I'm generally pro-acronym and not too sensitive about religion/spirituality. Second was this snap-shot of physician belief :
Of the respondents, 10% reported no religious affiliation, 59% reported being Christian, 16% reported being Jewish, and 14% reported other affiliations; 41% agreed with the statement, "My whole approach to life is based on my religion." Forty-one percent of the physicians believed it was appropriate for them to talk about their own R/S with patients when the patient asked about it. Fifty percent of the physicians believed it was appropriate for them to pray with patients when the patient requested it. The physicians reported that R/S "often" had a positive impact on their patients...and "rarely" had a negative impact.... 6)
I think I was so distracted by the vertebroplasty debacle in August that I forgot to mention another paper from the same article in the NEJM on prognosis after cardiac arrest. This is excellent teaching file material, particularly for fellows - while these guidelines remain the most widely used ones, this paper has a couple important additions to those. First, and speaking personally here, it has a welcome explanation of the whole somatosensory evoked potentials/N20 response test (an EEG/electrophysiologic test). This is something I've read about, seen my neurology colleagues employ it, but not really have any idea what it involved: this paper explains it (although does not speculate as to why an absent N20 response is such a dire prognostic indicator). Second, it discusses hypothermia protocols and what they might do to prognosis. No answers here, although there's reason to think that hypothermia may improve the general prognosis for those with anoxic ischemic encephalopathy and no brisk/early signs of recovery from absolutely wretched, to just very poor. The author did note that in a recent trial a few patients who had no-better-than-extensor motor responses on day 3 (which has been considered a near universal marker of near zero chance of meaningful recovery) improved by day 6 (showed signs of awareness). Given that many families decide to cease life-prolonging measures several days out from an arrest it's unclear how best to study this: including only those who live a while and whose families choose to continue life-prolonging treatments will overestimate survival and neurologic recovery, as inevitably these patients will be younger with fewer comorbidities than those who died after discontinuing life-sustaining treatments....
The American Journal of Speech-Language Pathology has a paper about SLP's attitudes and beliefs about feeding tubes for patients with advanced dementia. It comes from a survey of a national sample of SLPs (n=350, 57% response rate). The data are confusing, at least on one level. About 80% believe FT improved the nutritional status of patients with advanced dementia and eating problems; about half thought they improved functional status; only 40% thought they improved quality of life. About 40% thought they were the standard of care in their community, although 15% thought that should be the case. About 56% of SLP, given a case of a patient with advanced dementia and dysphagia, said they'd recommend a PEG tube placement.
However, when given a similar scenario, only 11% said they'd want their family to ask for a PEG tube to be placed in them.
This is a stark demonstration of the chasm between what most of us would want for ourselves if we developed advanced dementia (be comfortable, no tubes) and then what we (health professionals) actually do to our patients (for, yes, extraordinarily complex reasons, which includes struggling family members).
I'm beginning to keep my eye out for research about the differences between what people want to happen and what actually happens. Let me rephrase that since most of us want to live long and healthy lives and die relatively quickly after a brief illness. So defining that chasm is an easy one. What I mean is research which more directly looks at what agents in our health care system (whether clinicians, family members, or patients) say they want/they'd want for themselves, then end up getting/doing. One example - which I think is a real study that I didn't hallucinate but can't for the life of me find now (please comment if you know this study) - looks at proxy decision makers for patients on chronic ventilators (chronically critically ill patients - I don't think these were quadraplegics or patients with ALS), asking them why they chose to keep their loved one on a vent, whether they regret it, etc. This study also asked the proxy (family member) if something similar happened to them would they want their family to maintain them on a vent like this. Most said 'No' - although they were electing to do this same thing to someone they loved. It really struck me as emblematic for so much of what we see go wrong here, and underscores a theme that I've returned to on the blog multiple times - is this a 'problem' with families? with individual clinicians (if we just counseled people better, gave clearer/more coherent/stronger recommendations/owned these decisions more - a lot of this could be avoided)? or with The System?