Everyone's so concerned about the trial design without addressing the complexity of sedation and delirium in the ICU. First time I read the trial I couldn't understand the clinical equipoise for conducting the study.
The limiting factor for time to extubation in the ICU is far more complex than drug infusion. Secondly precedex can never achieve the RASS goal propofol can. You're comparing two unequal agents.
Finally, the amount of crossover in the trial made any serious data dissection moot. The only conclusion to draw is that if you need deep sedation propofol is more effective ---- but we already knew that.
Pragmatic trials CAN work when the intervention isnt messy. There's too much noise in the ICU thats why 99% of ICU trials are negative and the remaining have a fragility index in the single digits.
The one thing that would make me uncomfortable about using dexmedetomidine is that more patients in that arm became agitated, and this would make me concerned about long-term psychological effects well after being discharged from the ICU.
I am hardly an expert in the specific subject matter but I am a mathematician and can analyze data. If I understood the article correctly, the 3-way trial was between the default medication Propofol and the newer sedation medications which I will call D and C to save typing time! Dr. Mandrola highlights the open-label problem: in the course of the trial, clinicians were free to mix and match D and C, with Propofol. Yes, this creates a practical problem in coming to a meaningful conclusion, no doubt. However to me, the far bigger problem is that it seems that when there was mixing of infusions, the study (which I admit I have yet to read) failed to analyze the amounts of infusions mixed, timing of the different doses, durations, or other parameters that could be looked at in an attempt to reduce the noise factor. I would think that if the goal of the study was to specifically identify which of Propofol, D and C gives superior performance, then the pragmatic approach is not all that helpful. Yes, it may better guide clinicians in a general way, but was that the goal of the trial? From the study's title, it seems not. I would think that clinicians would want to start by having some insight into the performance comparisons between the 3, and with that done, then dig into to the greater complexities. The way this study was done, we have no insight into the signal-to-noise level of the conclusion.
When I was a med resident at SFGH in 1966 I was on the GI service and with the fellow we were involved in a trial of IV Valium for sedation during esophagoscopy. Lots of alcoholic GI bleeders looking for varices. For the trial Roche gave us ampules of Valium and placebo to use randomly (no sedation for some - that's how trials went in those days). The only problem was that the placebo was clear and the good stuff was yellow tinted. If a patient was really sick we used the clear ampules since we knew the yellow tinted stuff tended to make the patients apneic. Those of us who did the trial clearly learned something, but I was never involved in reporting the results, so I don't know what lesson the medical community learned.
I think the better question to ask here is how does no sedation compare to light station with any of these drugs. There is limited evidence about awake ICUs. Despite what you write in the article, it seems clear that not all patients NEED sedation to be intubated.
Generally a big fan of pragmatic trials. Yes there is noise but welcome to clinical medicine. I was moving more towards precedex/clonisine based upon a N=1 clinical impression but this study bolsters my confidence that propofol is an adequate substitute/ adjunct. Research benefiting the practicing doc- imagine that!
There are some interesting findings even in well controlled trials.
There were two trials that led to practice changes in time-dose-fractionation of post-lumpectomy breast radiation. The British START-B trial and the Canadian accelerated hypofractionated breast radiation.
Both trials established as non-inferior 15 fraction irradiation at higher dose per day over conventional fractionated at 30-33 treatment at lower doses. The difference was the Canadian (Whelan) trial which did not allow a series of 5 extra (boost) treatments to the lumpectomy cavity, while the British trial did at physician discretion.
Both trial outcomes were the same: the experimental arm was non-inferior. The controversy was the role of the boost dose, especially in elevated risk patients. My practice continued to use the boost after breast radiation was complete using the British trial as a rationale.
Ultimately Princess Margaret Hospital/Univ. Toronto Sunnybrook did a post study review of the Whelan trial looking at how their study patients were actually treated. They found that many of the patients did, in fact, receive a boost which was not permitted on the protocol. So, we really had no answer to the need for a boost in potentially curative breast cancer treatments with breast conserving therapy, except for a side finding that demonstrated the boost may be necessary in certain non-invasive breast cancer.
A second case was a controversy driven by M.D. Anderson concerning the need for post mastectomy radiation (PMRT) in node positive breast cancer. MDACC said only necessary for >4 nodes positive, while a Danish study 82B/C said there was a survival benefit in PMRT with any node positive disease. MDACC based their reasoning on the lower number of nodes taken in the EU in general. Overgaard took another look at their data in a planned subset analysis, selecting only patients with >10 nodes dissected, that her original data held: There is a benefit in PMRT with any nodal disease found.
Both of these trials led to significant improvements in standards of care. A careful consideration and effective response to questions about the study protocol and subsequent efforts to re-evaluate study criticism and questions led to greater information than would have been determined had these studies not been subject to further post-study evaluation based on real world clinical experience and questions.
Conclusion: clinical experience and intuition, even in a RCS trial has merit and benefit, provided the results of the trials are properly analyzed, even when that analysis was not part of the original pre-planned study parameters.
Also- in the icu we often are switching back and forth to different agents depending on tolerance, side effects (like bradycardia) etc. But I also would want to know how long someone was on the infusions.
Kids can’t be on propofol long bc of propofol infusion syndrome- metabolic complication. But propofol comes off easier than precedex when a patient has been on it a long time- and even then we often have to give doses of other medications to wean them off the dex. At least in kids there are many factors when selecting drugs.
Nothing is said here about the most important factor. What were the negative factors of using Propofol that led to the trial even being done? And were those negative factors measured in the different arms to see if there was improvement?
Everyone's so concerned about the trial design without addressing the complexity of sedation and delirium in the ICU. First time I read the trial I couldn't understand the clinical equipoise for conducting the study.
The limiting factor for time to extubation in the ICU is far more complex than drug infusion. Secondly precedex can never achieve the RASS goal propofol can. You're comparing two unequal agents.
Finally, the amount of crossover in the trial made any serious data dissection moot. The only conclusion to draw is that if you need deep sedation propofol is more effective ---- but we already knew that.
Pragmatic trials CAN work when the intervention isnt messy. There's too much noise in the ICU thats why 99% of ICU trials are negative and the remaining have a fragility index in the single digits.
The one thing that would make me uncomfortable about using dexmedetomidine is that more patients in that arm became agitated, and this would make me concerned about long-term psychological effects well after being discharged from the ICU.
What about measuring outcomes I add a patient an interested in - post ICU PTSD? Who wants to be alive with debilitating PTSD? Not me
I am hardly an expert in the specific subject matter but I am a mathematician and can analyze data. If I understood the article correctly, the 3-way trial was between the default medication Propofol and the newer sedation medications which I will call D and C to save typing time! Dr. Mandrola highlights the open-label problem: in the course of the trial, clinicians were free to mix and match D and C, with Propofol. Yes, this creates a practical problem in coming to a meaningful conclusion, no doubt. However to me, the far bigger problem is that it seems that when there was mixing of infusions, the study (which I admit I have yet to read) failed to analyze the amounts of infusions mixed, timing of the different doses, durations, or other parameters that could be looked at in an attempt to reduce the noise factor. I would think that if the goal of the study was to specifically identify which of Propofol, D and C gives superior performance, then the pragmatic approach is not all that helpful. Yes, it may better guide clinicians in a general way, but was that the goal of the trial? From the study's title, it seems not. I would think that clinicians would want to start by having some insight into the performance comparisons between the 3, and with that done, then dig into to the greater complexities. The way this study was done, we have no insight into the signal-to-noise level of the conclusion.
When I was a med resident at SFGH in 1966 I was on the GI service and with the fellow we were involved in a trial of IV Valium for sedation during esophagoscopy. Lots of alcoholic GI bleeders looking for varices. For the trial Roche gave us ampules of Valium and placebo to use randomly (no sedation for some - that's how trials went in those days). The only problem was that the placebo was clear and the good stuff was yellow tinted. If a patient was really sick we used the clear ampules since we knew the yellow tinted stuff tended to make the patients apneic. Those of us who did the trial clearly learned something, but I was never involved in reporting the results, so I don't know what lesson the medical community learned.
I think the better question to ask here is how does no sedation compare to light station with any of these drugs. There is limited evidence about awake ICUs. Despite what you write in the article, it seems clear that not all patients NEED sedation to be intubated.
Generally a big fan of pragmatic trials. Yes there is noise but welcome to clinical medicine. I was moving more towards precedex/clonisine based upon a N=1 clinical impression but this study bolsters my confidence that propofol is an adequate substitute/ adjunct. Research benefiting the practicing doc- imagine that!
Thank you Dr.
I just read an article from
F. Perry Wilson MD
RE treating AFib diagnosed by smartwatch. It was interesting. I wonder if you can comment on it.
Thank you
Frank
There are some interesting findings even in well controlled trials.
There were two trials that led to practice changes in time-dose-fractionation of post-lumpectomy breast radiation. The British START-B trial and the Canadian accelerated hypofractionated breast radiation.
Both trials established as non-inferior 15 fraction irradiation at higher dose per day over conventional fractionated at 30-33 treatment at lower doses. The difference was the Canadian (Whelan) trial which did not allow a series of 5 extra (boost) treatments to the lumpectomy cavity, while the British trial did at physician discretion.
Both trial outcomes were the same: the experimental arm was non-inferior. The controversy was the role of the boost dose, especially in elevated risk patients. My practice continued to use the boost after breast radiation was complete using the British trial as a rationale.
Ultimately Princess Margaret Hospital/Univ. Toronto Sunnybrook did a post study review of the Whelan trial looking at how their study patients were actually treated. They found that many of the patients did, in fact, receive a boost which was not permitted on the protocol. So, we really had no answer to the need for a boost in potentially curative breast cancer treatments with breast conserving therapy, except for a side finding that demonstrated the boost may be necessary in certain non-invasive breast cancer.
A second case was a controversy driven by M.D. Anderson concerning the need for post mastectomy radiation (PMRT) in node positive breast cancer. MDACC said only necessary for >4 nodes positive, while a Danish study 82B/C said there was a survival benefit in PMRT with any node positive disease. MDACC based their reasoning on the lower number of nodes taken in the EU in general. Overgaard took another look at their data in a planned subset analysis, selecting only patients with >10 nodes dissected, that her original data held: There is a benefit in PMRT with any nodal disease found.
Both of these trials led to significant improvements in standards of care. A careful consideration and effective response to questions about the study protocol and subsequent efforts to re-evaluate study criticism and questions led to greater information than would have been determined had these studies not been subject to further post-study evaluation based on real world clinical experience and questions.
Conclusion: clinical experience and intuition, even in a RCS trial has merit and benefit, provided the results of the trials are properly analyzed, even when that analysis was not part of the original pre-planned study parameters.
Also- in the icu we often are switching back and forth to different agents depending on tolerance, side effects (like bradycardia) etc. But I also would want to know how long someone was on the infusions.
Kids can’t be on propofol long bc of propofol infusion syndrome- metabolic complication. But propofol comes off easier than precedex when a patient has been on it a long time- and even then we often have to give doses of other medications to wean them off the dex. At least in kids there are many factors when selecting drugs.
Why not use a double dummy design to mask actual treatment assignment and limit crossover?
Does it matter at all how many and what drugs the patient is already taking?
Nothing is said here about the most important factor. What were the negative factors of using Propofol that led to the trial even being done? And were those negative factors measured in the different arms to see if there was improvement?