Thank you! This is a great (and overdue) reminder that “GDMT for everyone, immediately” can turn evidence-based medicine into checkbox medicine when we ignore external validity. Trials like PARADIGM-HF weren’t just testing drugs; they were testing drugs in a high-support environment with run-ins, careful selection, and monitoring that systematically excludes (or filters out) many of the patients we actually see: the frail 80-year-old with borderline BP, CKD, polypharmacy, limited transportation, and no “trial nurse navigator” to keep the regimen on the rails. What I appreciated most is the call for minimally disruptive care: for robust ambulatory HFrEF patients, full-court-press GDMT is often exactly right. But for multimorbid, vulnerable patients labeled “HF”, the highest-quality care may be prioritization (diuresis, a tolerable low-dose neurohormonal backbone, de-prescribing, cost-aware choices, and goals-of-care alignment) without being punished by metrics that confuse “more meds” with “better care”. If we want quality measures that truly improve outcomes, they should reward appropriate intensification and appropriate restraint, because in real-world heart failure, the art is not knowing the four pillars; it’s knowing when and how they’re safely deployable.
Real-world data shows that many patients with heart failure in community settings don't receive the full guideline-directed medical therapy (GDMT) due to practical barriers like lack of access and financial constraints.
To Mattman26: Review the following studies/trials: SOLVD; Val-Heft; Valiant; Elite-II; CHARM. ARB’s reduce hospital admissions and can improve quality of life but have failed in terms of reduction in all cause mortality. They are recomended for patients that don’t tolerated ACE-I. For this reason, the PARADIGM Trial compared SAVA with Enalapril(showed benefit reducing death due to all causes).
This lines up with my concern about population-level guidelines for screenings, etc. We need to treat the person in front of us instead of simply employing the latest guideline/protocol/bundle without genuinely assessing if it beneficial for the person in front of us.
Let’s be careful! GDMT has been reported as sub-optimally used in Europe and the USA. In countries in South America, GDMT is “named” as “CRT of the poor”(lack of resources for CRM devices and lower cost of medications due to their “non-domestic” status makes attaining GDMT almost compelling). Positive remodeling is achieved significantly “better” than in Europe and USA. We must continue to encourage the use of GDMT. Of course, close follow up including eGFR, K levels, BP, patient’s clinical response needs for good clinicians to be in charge. In the study you made reference published in November, 2025 the statement that all patients were in similar neurohormonal drug regimens means currently recommended GDMT? Let’s use common sense combined with knowledge (LOGIC). Frailty can mislead clinicians reading your post and some suitable patients could probably be denied therapies that can improve their quality of life and survival. Thanks and happy holidays!!!
It is ironic that in the name of EBM, an **observational** uncontrolled study (non-HF is not a control) is being used to argue against a **controlled RCT**, in this case, the argument is it is too controlled.
The observational study wasn't even designed to measure the relative benefit of new vs. old drugs.
Imagine an **observational** study showed benefit of newer drugs, I can see Dr. Mandrola will be arguing why it is useless because it is not controlled.
If the RCT is not generalizable to the patient in front of you, better to reference a relevant observational study than an RCT with no external validity. Critical appraisal, a core component of EBM, must first be done to see if the study is even applicable to the clinical scenario before you.
This is where Dr. Mandrola is using 2 different standards. He holds his own biases to a much lower level of evidence.
Dismissing HF trials not including older patients without making an argument why human biology stops working once you reach 70 years old. It is a red herring.
I can also argue that these controlled RCTs have not shown that the drug(s) work for women with purple hair.
To supports his own bias, Dr. Mandrola uses an uncontrolled observational study that was NOT designed to measure the the benefit of medication regimen A and B. It is merely measuring mortality trend of HF diagnosed patients by 5 year cohort. He is arguing that there was no separation of mortality curves between 2011-15 and 2016-20 cohorts (purple and green lines in figure 2).
For one of your next posts maybe it would be worth while considering to do something similar about the valvular guidelines
The Coapt trial which despite the contrarian results of MITRA FR has hugely influenced current recommendations for TEER in secondary MR has took almost 10 years to enrol patients in topgun world Centers
SGLT2is are the most annoying. Cardiologists have a month or two of samples. The patient likes the meds; we adjust their diabetes regimen so they don't become hypoglycemic; the samples end; the copay is a ridiculous amount of money; the patient wants us to explore all the options to get it affordably; we can't find an affordable option; we revert their diabetes regimen. It's not the best use of anyone's time, not the safest thing for the patients, or the most effective thing for the patients. I would like a trial where patients go on and off the meds, screw up their stable diabetes regimen, have an episode of hypo- or hyper- glycemia or two, and because we spent the time screwing around with the SGLT2i have some other need not met or maybe just have to find a new primary care doctor earlier than expected because she ran screaming off into the woods, unable to cope with yet another unexpected medication change.
Everyone’s insurance is different and the pharmacy cannot check the cost without the doctor sending a prescription. Then someone has to call the pharmacy and wade thru phone tree hell to find out the cost and tell the patient who is long gone and either took the samples with them (which the cardiologist now has to call and say ‘don’t take’ and which have been wasted) or didn’t take the samples and now the patient’s child (because the patient doesn’t drive to the town where the specialist is) has to come pick the meds up and by the time they’ve run out of samples, their insurance has changed how much the patient has to pay for that med and now it’s unaffordable for them.
And I’m not even exaggerating except maybe the part about the patient’s kid driving. Sometimes it’s their grandchildren or their neighbor who drives in to pick up the med.
“Thought leaders” (esp in the HF space) are no longer “thinkers” of any desirable repute. It’s 4 drugs for everyone, all day everyday, regardless of the nuances even within the foundational studies themselves. Never mind the patient characteristics that make most patients you see in clinic different from those enrolled in those foundational studies.
So when a patient resembles those in the trials, by all means apply that data. But one should recognize when they don’t.
This penchant for “knowledge translation” has merely resulted in the dumbing down of the profession, and overuse of the products of Pharma.
I often return to Sackett. The “evidence” is only one sphere within the EBM paradigm. Those who ignore unique patient characteristics/comorbidities (“clinical experience”) and/or patient values and preferences are NOT practicing EBM in any rigorous form.
As an HF patient, I have never received a satisfactory answer (emphasis on satisfactory) as to how we can conclude that the addition of sacubatril to old-school Valsartan is helpful when they didn’t compare it to Valsartan without the addition of sacubatril. Vinay has been all over this. It’s nuts.
Without well-run, large-scale RCTs, we would have no idea if a drug or approach or device works consistently better than current approaches. But these are studies. That is, they are experiments. To get the clearest answer to the experimental question, strict entry criteria ensure that patients are similar enough in the most relevant ways, and the investigators hope that all other differences among patients are randomly distributed between the control and experimental arms. If 20% of your large group of patients can’t tolerate the drug, it doesn’t make sense to include them in a seminal study. Otherwise, we’d never learn anything about whether our drug works when it’s given its best chance to do so. It’s how good science makes progress.
But as John so clearly points out, the findings of a large RCT are only applicable to the kinds of patients who made it through the study. The guideline makers who confuse experiments with real life would do well to remember the difference.
I'm still another retired family doc. I guess this is the place we like to hang out.
Two things come to mind. One is not only is the HF population in a community setting different than in an academic study, when I was still seeing hospital patients we were encouraged to assign every diagnosis possible in order to increase reimbursement. If there was any hint of HF in the chart there would be a reviewer coming by to encourage you to add that diagnosis to the current hospitalization. That then gets them into the QA protocols of course.
The second is something I learned early in my career - in order for the conclusions of a study to be applied to a given patient, the conditions of that study have to be replicated. I think that is saying the same thing Dr Mandrola is saying in a different way. The academic studies are being applied to patients in much different settings, and the study conclusions may or may not be valid.
Thank you! This is a great (and overdue) reminder that “GDMT for everyone, immediately” can turn evidence-based medicine into checkbox medicine when we ignore external validity. Trials like PARADIGM-HF weren’t just testing drugs; they were testing drugs in a high-support environment with run-ins, careful selection, and monitoring that systematically excludes (or filters out) many of the patients we actually see: the frail 80-year-old with borderline BP, CKD, polypharmacy, limited transportation, and no “trial nurse navigator” to keep the regimen on the rails. What I appreciated most is the call for minimally disruptive care: for robust ambulatory HFrEF patients, full-court-press GDMT is often exactly right. But for multimorbid, vulnerable patients labeled “HF”, the highest-quality care may be prioritization (diuresis, a tolerable low-dose neurohormonal backbone, de-prescribing, cost-aware choices, and goals-of-care alignment) without being punished by metrics that confuse “more meds” with “better care”. If we want quality measures that truly improve outcomes, they should reward appropriate intensification and appropriate restraint, because in real-world heart failure, the art is not knowing the four pillars; it’s knowing when and how they’re safely deployable.
Real-world data shows that many patients with heart failure in community settings don't receive the full guideline-directed medical therapy (GDMT) due to practical barriers like lack of access and financial constraints.
I completely agree with Dr Mandrola.
Besides, we have to take into account that there is no evidence for using the 4 drugs almost at the same time in all patients with heart failure.
This is the artcile that theoretically supports the use of 4 drugs:
Estimating lifetime benefits of comprehensive
disease-modifying pharmacological therapies in patients
with heart failure with reduced ejection fraction:
a comparative analysis of three randomised controlled trials
Muthiah Vaduganathan, Brian L Claggett, Pardeep S Jhund, Jonathan W Cunningham, João Pedro Ferreira, Faiez Zannad, Milton Packer,
Gregg C Fonarow, John J V McMurray, Scott D Solomon.
I recommend to read it and analyze it, because we are dealing with a mathematical model based in Paradigm, Emphasis an Dapa hf.
The finding was hypothesis generating. Not a clear truth.
To date there is no trial that compares 4 drugs vs beta blockers plus ACEI in all patients with heart failure.
Do I have to give a leap fo faith?
Excellent post!!!
To Mattman26: Review the following studies/trials: SOLVD; Val-Heft; Valiant; Elite-II; CHARM. ARB’s reduce hospital admissions and can improve quality of life but have failed in terms of reduction in all cause mortality. They are recomended for patients that don’t tolerated ACE-I. For this reason, the PARADIGM Trial compared SAVA with Enalapril(showed benefit reducing death due to all causes).
This lines up with my concern about population-level guidelines for screenings, etc. We need to treat the person in front of us instead of simply employing the latest guideline/protocol/bundle without genuinely assessing if it beneficial for the person in front of us.
Let’s be careful! GDMT has been reported as sub-optimally used in Europe and the USA. In countries in South America, GDMT is “named” as “CRT of the poor”(lack of resources for CRM devices and lower cost of medications due to their “non-domestic” status makes attaining GDMT almost compelling). Positive remodeling is achieved significantly “better” than in Europe and USA. We must continue to encourage the use of GDMT. Of course, close follow up including eGFR, K levels, BP, patient’s clinical response needs for good clinicians to be in charge. In the study you made reference published in November, 2025 the statement that all patients were in similar neurohormonal drug regimens means currently recommended GDMT? Let’s use common sense combined with knowledge (LOGIC). Frailty can mislead clinicians reading your post and some suitable patients could probably be denied therapies that can improve their quality of life and survival. Thanks and happy holidays!!!
It is ironic that in the name of EBM, an **observational** uncontrolled study (non-HF is not a control) is being used to argue against a **controlled RCT**, in this case, the argument is it is too controlled.
The observational study wasn't even designed to measure the relative benefit of new vs. old drugs.
Imagine an **observational** study showed benefit of newer drugs, I can see Dr. Mandrola will be arguing why it is useless because it is not controlled.
If the RCT is not generalizable to the patient in front of you, better to reference a relevant observational study than an RCT with no external validity. Critical appraisal, a core component of EBM, must first be done to see if the study is even applicable to the clinical scenario before you.
This is where Dr. Mandrola is using 2 different standards. He holds his own biases to a much lower level of evidence.
Dismissing HF trials not including older patients without making an argument why human biology stops working once you reach 70 years old. It is a red herring.
I can also argue that these controlled RCTs have not shown that the drug(s) work for women with purple hair.
To supports his own bias, Dr. Mandrola uses an uncontrolled observational study that was NOT designed to measure the the benefit of medication regimen A and B. It is merely measuring mortality trend of HF diagnosed patients by 5 year cohort. He is arguing that there was no separation of mortality curves between 2011-15 and 2016-20 cohorts (purple and green lines in figure 2).
Please read the paper here: https://www.thelancet.com/journals/lanhl/article/PIIS2666-7568(24)00029-1/fulltext
In fact, SGLT2i usage was so low that it would not have changed mortality rates because so few patients used them. 0.1% in 2011-15 and 2% in 2016-20.
So this was the "evidence" used to argue that new drugs don't work in patients who were not in the RCT cohorts.
Do you see the contradiction here? Dr. Mandrola argues for his biases using much lower quality evidence.
Can you explain? I don't follow
Great week. So sad when my patients feel that they are doing less than because they can’t afford enTresto.
Great Post John
For one of your next posts maybe it would be worth while considering to do something similar about the valvular guidelines
The Coapt trial which despite the contrarian results of MITRA FR has hugely influenced current recommendations for TEER in secondary MR has took almost 10 years to enrol patients in topgun world Centers
SGLT2is are the most annoying. Cardiologists have a month or two of samples. The patient likes the meds; we adjust their diabetes regimen so they don't become hypoglycemic; the samples end; the copay is a ridiculous amount of money; the patient wants us to explore all the options to get it affordably; we can't find an affordable option; we revert their diabetes regimen. It's not the best use of anyone's time, not the safest thing for the patients, or the most effective thing for the patients. I would like a trial where patients go on and off the meds, screw up their stable diabetes regimen, have an episode of hypo- or hyper- glycemia or two, and because we spent the time screwing around with the SGLT2i have some other need not met or maybe just have to find a new primary care doctor earlier than expected because she ran screaming off into the woods, unable to cope with yet another unexpected medication change.
Maybe reconsider starting a patient on samples of a drug you know they can't afford?
Everyone’s insurance is different and the pharmacy cannot check the cost without the doctor sending a prescription. Then someone has to call the pharmacy and wade thru phone tree hell to find out the cost and tell the patient who is long gone and either took the samples with them (which the cardiologist now has to call and say ‘don’t take’ and which have been wasted) or didn’t take the samples and now the patient’s child (because the patient doesn’t drive to the town where the specialist is) has to come pick the meds up and by the time they’ve run out of samples, their insurance has changed how much the patient has to pay for that med and now it’s unaffordable for them.
And I’m not even exaggerating except maybe the part about the patient’s kid driving. Sometimes it’s their grandchildren or their neighbor who drives in to pick up the med.
Very informative, thank you.
That was no misunderstanding.
“Thought leaders” (esp in the HF space) are no longer “thinkers” of any desirable repute. It’s 4 drugs for everyone, all day everyday, regardless of the nuances even within the foundational studies themselves. Never mind the patient characteristics that make most patients you see in clinic different from those enrolled in those foundational studies.
So when a patient resembles those in the trials, by all means apply that data. But one should recognize when they don’t.
This penchant for “knowledge translation” has merely resulted in the dumbing down of the profession, and overuse of the products of Pharma.
I often return to Sackett. The “evidence” is only one sphere within the EBM paradigm. Those who ignore unique patient characteristics/comorbidities (“clinical experience”) and/or patient values and preferences are NOT practicing EBM in any rigorous form.
As an HF patient, I have never received a satisfactory answer (emphasis on satisfactory) as to how we can conclude that the addition of sacubatril to old-school Valsartan is helpful when they didn’t compare it to Valsartan without the addition of sacubatril. Vinay has been all over this. It’s nuts.
Without well-run, large-scale RCTs, we would have no idea if a drug or approach or device works consistently better than current approaches. But these are studies. That is, they are experiments. To get the clearest answer to the experimental question, strict entry criteria ensure that patients are similar enough in the most relevant ways, and the investigators hope that all other differences among patients are randomly distributed between the control and experimental arms. If 20% of your large group of patients can’t tolerate the drug, it doesn’t make sense to include them in a seminal study. Otherwise, we’d never learn anything about whether our drug works when it’s given its best chance to do so. It’s how good science makes progress.
But as John so clearly points out, the findings of a large RCT are only applicable to the kinds of patients who made it through the study. The guideline makers who confuse experiments with real life would do well to remember the difference.
I'm still another retired family doc. I guess this is the place we like to hang out.
Two things come to mind. One is not only is the HF population in a community setting different than in an academic study, when I was still seeing hospital patients we were encouraged to assign every diagnosis possible in order to increase reimbursement. If there was any hint of HF in the chart there would be a reviewer coming by to encourage you to add that diagnosis to the current hospitalization. That then gets them into the QA protocols of course.
The second is something I learned early in my career - in order for the conclusions of a study to be applied to a given patient, the conditions of that study have to be replicated. I think that is saying the same thing Dr Mandrola is saying in a different way. The academic studies are being applied to patients in much different settings, and the study conclusions may or may not be valid.