Friday, May 2, 2008
Evidence and palliative care; hydration reviewed, sort of; opioid antagonists for OBD
Over here at Pallimed we’ve been having an off-line conversation about “evidence” for palliative care-related practices, and decided to bring it to our readers. The impetus for the discussion is the publication of two new Cochrane reviews, one on medical hydration in advanced illness and the other on mu-opioid antagonists for opioid-induced bowel dysfunction.
Regular readers of Pallimed will know that the three of us are strong proponents of increasing and improving the evidence base for palliative care. We are also very much aware of the difficulty of doing so, especially using the stricter definitions and methods of evidence-based medicine (EBM).
I will disclose that I am not a regular devotee of the Cochrane reviews, but “medically assisted hydration” caught my eye. I was disheartened, however to see that only 5 papers made the cut to be reviewed, and that the conclusion was “There are insufficient good quality studies to make any recommendations for practice with regard to the use of medically assisted hydration in palliative care patients.” It reminded me of a conversation a couple of years ago with a well-known pain physician and researcher. He told me that he doesn't even read Cochrane reviews anymore. "They all end with the same conclusion: there is insufficient evidence to make a recommendation about . . . " I don't think that is necessarily a bad thing, but they are super-strict in their criteria for selection. That makes it particularly difficult in palliative care, where doing the type of research one might do for approving a hypertension agent is almost impossible. The issue, of course, is that the assumed “highest” level of evidence is the randomized controlled trial (RCT). Some have argued that RCTs are not only difficult to design and conduct, but may be unethical in interventions for people with advanced disease where comfort, not cure or control, is the therapeutic goal. Others criticize the misuse of EBM, either attempting to apply broadly the results from a trial in a subset if patients, or conversely denying payment for a successful intervention for a specific individual because the RCT “proves” that the intervention doesn’t work. Fragile patients with advanced disease are usually excluded from RCTs, yet that is precisely the population most in need of evidence-based palliative interventions.
The phrase “perfect is the enemy of good” springs to mind. But the RCT is unlikely to ever be the “perfect” tool for symptom management, population-based or public health studies, and other complex beyond-the-physiology questions.
Carr, an early proponent, points out that EBM continues to evolve, has limits, and can easily be misused. Where applicable, the RCT should be used, but with the results interpreted judiciously for specific populations and individuals. Hallenbeck, Carr (specifically for pain medicine), Aoun & Kristjanson, and Devery are among those who have made reasoned criticisms of EBM, and suggested “other ways” of using EBM or of gaining knowledge. These include an “Equity-based evidence framework” (Aoun & Kristjanson), “narrative-based” evidence (Devery), or recognition that “lower” (not necessarily less rigorous) levels of evidence apply quite well to palliative care.
We’d like to hear how our readers use and interpret “evidence” in their practice of palliative care.
2)
Regular readers of Pallimed will know that the three of us are strong proponents of increasing and improving the evidence base for palliative care. We are also very much aware of the difficulty of doing so, especially using the stricter definitions and methods of evidence-based medicine (EBM).
I will disclose that I am not a regular devotee of the Cochrane reviews, but “medically assisted hydration” caught my eye. I was disheartened, however to see that only 5 papers made the cut to be reviewed, and that the conclusion was “There are insufficient good quality studies to make any recommendations for practice with regard to the use of medically assisted hydration in palliative care patients.” It reminded me of a conversation a couple of years ago with a well-known pain physician and researcher. He told me that he doesn't even read Cochrane reviews anymore. "They all end with the same conclusion: there is insufficient evidence to make a recommendation about . . . " I don't think that is necessarily a bad thing, but they are super-strict in their criteria for selection. That makes it particularly difficult in palliative care, where doing the type of research one might do for approving a hypertension agent is almost impossible. The issue, of course, is that the assumed “highest” level of evidence is the randomized controlled trial (RCT). Some have argued that RCTs are not only difficult to design and conduct, but may be unethical in interventions for people with advanced disease where comfort, not cure or control, is the therapeutic goal. Others criticize the misuse of EBM, either attempting to apply broadly the results from a trial in a subset if patients, or conversely denying payment for a successful intervention for a specific individual because the RCT “proves” that the intervention doesn’t work. Fragile patients with advanced disease are usually excluded from RCTs, yet that is precisely the population most in need of evidence-based palliative interventions.
The phrase “perfect is the enemy of good” springs to mind. But the RCT is unlikely to ever be the “perfect” tool for symptom management, population-based or public health studies, and other complex beyond-the-physiology questions.
Carr, an early proponent, points out that EBM continues to evolve, has limits, and can easily be misused. Where applicable, the RCT should be used, but with the results interpreted judiciously for specific populations and individuals. Hallenbeck, Carr (specifically for pain medicine), Aoun & Kristjanson, and Devery are among those who have made reasoned criticisms of EBM, and suggested “other ways” of using EBM or of gaining knowledge. These include an “Equity-based evidence framework” (Aoun & Kristjanson), “narrative-based” evidence (Devery), or recognition that “lower” (not necessarily less rigorous) levels of evidence apply quite well to palliative care.
We’d like to hear how our readers use and interpret “evidence” in their practice of palliative care.
2)
Getting back to the Cochrane reviews:
Hydration: there was only one “high quality” study (2 RCTs) but over a very short duration (2 days) and very underpowered. Overall, results of the 5 studies reviewed were somewhat contradictory, but showed suggestions of improvement in sedation and myoclonus. Negative effects cited were fluid retention leading to peripheral edema, ascites, and pleural effusion. In their conclusion, the authors acknowledge the difficulty of conducting clinical research in the palliative care population. They also comment “the issue of medically assisted hydration in palliative care patients causes such divergent views, yet there are so few studies to guide clinical practice properly. As well as looking at further RCTs in this area, the evidence base will be improved with at least more prospective controlled trials.” It should be noted that the "best" study had quite a few patients with reasonably good performance status, especially in the intervention group. It was not really helpful, therefore, in answering the question: "is medically assisted hydration a helpful intervention in patients in the last days of life?" One of the included studies, a prospective controlled trial, attempted to mimic real world decision-making by allowing physician preference in allocating to the intervention arm. This introduces bias,of course, but downgrading the score on the study design because of it may make some clinicians grind their teeth. The whole point in reading a study is to find help in making real world decisions, isn't it? That raises questions for a future conversation.
Opioid antagonists for opioid-induced bowel dysfunction (OBD):
Naloxone, nalbuphine, methylnaltrexone, and alvimopan were reviewed. There was only one study of nalbuphine. All studies were placebo-controlled RCTs (which makes sense in this case; active-controlled studies would also be welcome). Some studies of the newer agents (methylnaltrexone & alvimopan) were in healthy volunteers. The authors report that their task was made more difficult by use of various opioids in trials, and by inconsistent definitions of postoperative ileus. In general, though, they found (in a limited number of studies with small numbers) that methylnaltrexone and alvimopan were sufficiently safe and effective (increased transit time/decreased constipation) to be labeled “promising.” It should be noted that methylnaltrexone was administered parenterally in these trials, and that oral methylnaltrexone has yet to be reviewed. Alvimopan trials were interrupted because of excess cardiac events. The current target indication for alvimopan (not yet approved) is postoperative ileus.
Hydration: there was only one “high quality” study (2 RCTs) but over a very short duration (2 days) and very underpowered. Overall, results of the 5 studies reviewed were somewhat contradictory, but showed suggestions of improvement in sedation and myoclonus. Negative effects cited were fluid retention leading to peripheral edema, ascites, and pleural effusion. In their conclusion, the authors acknowledge the difficulty of conducting clinical research in the palliative care population. They also comment “the issue of medically assisted hydration in palliative care patients causes such divergent views, yet there are so few studies to guide clinical practice properly. As well as looking at further RCTs in this area, the evidence base will be improved with at least more prospective controlled trials.” It should be noted that the "best" study had quite a few patients with reasonably good performance status, especially in the intervention group. It was not really helpful, therefore, in answering the question: "is medically assisted hydration a helpful intervention in patients in the last days of life?" One of the included studies, a prospective controlled trial, attempted to mimic real world decision-making by allowing physician preference in allocating to the intervention arm. This introduces bias,of course, but downgrading the score on the study design because of it may make some clinicians grind their teeth. The whole point in reading a study is to find help in making real world decisions, isn't it? That raises questions for a future conversation.
Opioid antagonists for opioid-induced bowel dysfunction (OBD):
Naloxone, nalbuphine, methylnaltrexone, and alvimopan were reviewed. There was only one study of nalbuphine. All studies were placebo-controlled RCTs (which makes sense in this case; active-controlled studies would also be welcome). Some studies of the newer agents (methylnaltrexone & alvimopan) were in healthy volunteers. The authors report that their task was made more difficult by use of various opioids in trials, and by inconsistent definitions of postoperative ileus. In general, though, they found (in a limited number of studies with small numbers) that methylnaltrexone and alvimopan were sufficiently safe and effective (increased transit time/decreased constipation) to be labeled “promising.” It should be noted that methylnaltrexone was administered parenterally in these trials, and that oral methylnaltrexone has yet to be reviewed. Alvimopan trials were interrupted because of excess cardiac events. The current target indication for alvimopan (not yet approved) is postoperative ileus.
.
Just as an aside, the last author on this review is Dan Carr, cited above in the EBM discussion.
There is another recent review here.
A just-published RCT was not part of either review. Subcutaneous methylnaltrexone (Relistor) was approved by the FDA last week for patients in late-stage advanced illness and should be available next month.
There is another recent review here.
A just-published RCT was not part of either review. Subcutaneous methylnaltrexone (Relistor) was approved by the FDA last week for patients in late-stage advanced illness and should be available next month.
.
So, who among our readers is waiting impatiently for methylnaltrexone to become available? Given the route, cost, and limited research, are you eager to give it a try? In general, are we adequately treating opioid-induced constipation, but need this for backup? How will you determine which patients receive it? How many other interventions do you need to go through before you determine that constipation is refractory to more conventional treatment? Oh, and do any of you actually use the newly minted term 'opioid bowel dysfunction' (OBD)? The Comments link awaits you.