Please note that the text below is not a full transcript and has not been copyedited. For more insight and commentary on these stories, subscribe to the This Week in Cardiology podcast, download the Medscape app or subscribe on Apple Podcasts, Spotify, or your preferred podcast provider. This podcast is intended for healthcare professionals only.
In This Week’s Podcast
For the week ending June 28, 2024, John Mandrola, MD, comments on the following news and features stories: Screening echocardiography, multivitamins, wasteful research, another pulmonary artery (PA) sensor for heart failure (HF), vascular closure devices (VCDs), and GLP-1 marketing as science research.
TWIC takes next week off to celebrate America’s Independence Day.
Screening for Valvular HD in the Elderly
The European Heart Journal - Cardiovascular Imaging has published a paper on screening that deserves mention. It’s funny how I am developing allergies as I age. If there is one thing in Medicine that is most likely to give me a rash, it is disease creation or medicalization of life. It’s why I despise labels such as stage A HF, pre-hypertension, and pre-diabetes. Being stage A or pre-something is equivalent to getting older and not dying early.
This study chronicles what happens if you invite older patients to have echocardiograms. First author Vasiliki Tsampasian and colleagues mostly from Norwich medical school in the United Kingdom invited about 10,000 older people who had no indication for echocardiography to have echocardiography.
Slightly more than half agreed to have their heart imaged.
The most common types of valvular heart disease (VHD) were regurgitation of the tricuspid (13.8%), mitral (12.8%), and aortic (8.3%) valves. The authors did not include trivial regurgitation.
They noted clinically significant VHD in 2.4% (2.2% moderate and 0.2% severe), with mitral and aortic regurgitation being the most common.
The only characteristic associated with significant VHD was age.
The authors tell us that the number needed to scan to diagnose one clinically significant case of VHD is 42 for individuals ≥ 60 years old and 15 for those ≥75 years old.
The average age of patients was 70 years. Slightly more than half were female.
While the top-line findings were that VHD had a pretty high prevenance, severe disease was found in only seven of more than 4000 patients.
Comments. I am sorry to be critical of this paper. I mean no malice to the authors who were quite balanced and descriptive in the manuscript. They tell us what they found. They write that further studies are needed. Kudos to them and the editors.
I would argue that further studies are not needed. Nearly every attempt at screening for serious disease fails to show a mortality benefit. Even when we screen for diseases like cancer, and vascular disease, we fail to show mortality benefits. Maybe selective aneurysm screening has some promise, but this in highly selected patients.
Screening old people for VHD is like screening for getting wet walking in the rain. VHD is part of aging. If you don’t die at a young age, you stand a very good chance of having an abnormal sonogram.
In this study, seven of 4000 patients had severe valvular heart disease discovered. But how many were labeled with a disease and underwent unnecessary testing? How many developed anxiety about their hearts?
The most common condition of aging discovered was tricuspid regurgitation. Well, it turns out that — at least in the United States — we now have two percutaneous treatments for that. We will soon have direct-to-consumer ads and posters in lobbies of hospitals with messages to ask your doctor about your tricuspid valve.
I realize that the British healthcare system doesn’t have the profit incentives of the American system, but I want to infuse any listener of this podcast with a bit of fear of imaging older people who complain of nothing. Please, friends, resist that urge.
The main job of doctors who care for older people is to not mess up their proven success. Less is almost always more in this group. Ordering tests in older people should scare you.
Vitamins and Wasteful Research
JAMA Network Open has published an observational study on multivitamin use. Once again, I mean no malice to the authors of this study, but I find this large undertaking problematic, even though 86 news outlets reported on it.
The authors begin by telling us that one in three Americans take a multivitamin. They also tell us that the US Preventive Services Task Force has reviewed many randomized controlled trials (RCTs), emphasis on randomized, and found no benefit on important health outcomes vs placebo.
If an area has nine RCTs, why would a non-random retrospective comparison study be helpful? The authors write that it is external validity and short follow-up. Well, the Physician’s Health Study II had 14,000 patients and a decade of follow-up and was negative.
The authors used three large population cohorts and basically correlated self-reported multivitamin use with death. They adjusted for confounding variables.
The overall study included nearly 400,000 patients, and they found no association between multivitamin use and mortality.
Comments. American taxpayers funded the study as it was supported by an National Institutes of Health grant. I briefly discuss this study for a number of reasons.
First, I put almost no weight on the paper’s findings. We already have multiple clinical trials showing no benefits from multivitamin use. If this observational study were positive, it would not be useful, because of likely biases in non-random comparisons.
The second reason I put no weight on the findings stems from the concept of analytic flexibility which I discussed on the May 10th episode and wrote about on May 29th.
In the paper by Dena Zeraatkar from McMaster University, her team showed that there are literally a quadrillion different analytic choices when doing large observational comparison studies, and that different analytic choices lead to different findings — some significantly positive, some significantly negative, and some null.
The authors of the multivitamin paper chose one analytic method, but there are many others, so who knows what others may have led to?
The other reason to cover this paper is to highlight the notion of wasteful research. There is no reason to conduct such a study. It wasted money. I am sorry to say that, but I don’t what else to say.
We don’t need more multivitamin studies, or vitamin D studies. Of all the problems we have in health, and healthcare, people taking a multivitamin does not rank highly. My approach to vitamin taking is to spend as little time on it as possible. If asked, I say there is no proven benefit, but if you want to have colored urine, it’s fine. Now let’s talk about your body mass index, or your blood pressure (BP), or your anticoagulation decision.
FDA Has Approved Another PA Sensor
We had CardioMems PA sensor for the management of patients with HF. This month, the US Food and Drug Administration (FDA) approved one called the Cordella PA sensor.
The company says its advantages are that patients can use it sitting whereas CardioMEMS has to be used supine. It also provides more “physiologic data.”
FDA says approval was supported by a study called PROACTIVE-HF. That sentence sounds good, doesn’t it. The device has a study approving it.
The problem is that the industry-sponsored trial wasn’t really a trial. It started as a trial, but there was this line in the slide deck: ”Following GUIDE HF, with FDA input, PROACTIVE-HF was changed from a randomized single-blind design to a single-arm open label design with blinded endpoint assessment and pre-specified safety and effectiveness endpoints defined via historical measures from previous trials.”
Basically, patients who had this device had to have less than 0.43 events per patient per 6 months. This was pre-determined from previous CARDIOMEMS trials. No comparisons with active controls.
Of course, the device met that goal. Every patient and every doctor knew the patient had a monitor. Researchers knew the goal was to not have a HF hospitalization. Nearly half the patients had a normal ejection fraction; 58% were taking an SGLT2 inhibitor and nearly 100% were on a loop diuretic.
I am not sure what is going on at the FDA, but this is nuts. Here was an opportunity to study a second-generation PA monitor, and FDA let it go.
I remind you that the null GUIDE HF trial was one of the biggest black marks in recent cardiology history.
Briefly, GUIDE HF screened 1500 patients, randomized 1000. All patients got the CardioMEMS device, but the PA data from control patients was not used.
The primary endpoint was all-cause death and HF events which could be a hospitalization for HF or urgent HF visit.
The event rate was 0.56 events per patient year in the active arm vs 0.64 events per patient year in the control arm.
The hazard ratio (HR) was 0.88 but the 95% confidence interval (CI) went from 0.74 to1.05 and the P-value was 0.16.
This was a negative result. It did not meet statistical significance. Did they conclude that the device did not provide a benefit?
No. Not at all. They then did a COVID-19 analysis looking at events pre COVID and during COVID. Boom. They got a positive P-value.
In the pre-COVID group, the HR was 0.81, CI went from 0.66 to 1.00 and the P-value was 0.049.
Not only is this an extremely dubious maneuver, but during COVID, a time when wireless outpatient management would likely excel, the benefit of the device seemed to vanish. Here the event rates were actually a bit higher in CardioMEMs arm.
FDA ignored these findings and expanded the indication for this device. And now, FDA is lowering the bar for second-generation devices based on an actually non-significant GUIDE HF trial.
I love cardiology. I think we do a lot of good for people. But we sure waste a lot of money on unproven technology. How can we be so susceptible?
Another aspect of this whole PA monitoring story is that there is a remote monitoring charge every 30 days, much like ILRs. Patients get bills every 30 days. So not only is there a huge upfront cost of the device, but there is now a revenue stream every 30 days.
Speaking of Wasteful Healthcare, Let’s Talk about Vascular Closure Devices in Electrophysiology.
When we do an atrial fibrillation (AF) ablation, we use multiple large-bore venous access points. The range is two to four access catheters in the femoral veins. We do AF ablation with patients on full anticoagulation, so getting hemostasis can be a challenge.
In the old days, a nurse or technician held pressure on the access points until the bleeding stopped. My friend Piotr Futyma from Poland taught me to use a suture and stopcock to facilitate hemostasis. It takes 45 seconds, and costs less than $5 for the suture and stopcock. It’s a figure eight stitch that stays in for a few hours. On the elegance scale it is a 10 out of 10.
Lo and behold, industry has sold some of my colleagues on various devices that go into the vein and occlude it. A number of companies make such devices.
I’ve tried them, and they work pretty well. Their selling point are that they work, and patients have an hour or two shorter bed rest.
They cost up to $500 per device. So, if you use four access points, that’s an extra $2000 per ablation. If your center does 500 ablations per year, you have added an extra million dollars to healthcare costs.
I know of no trials of these devices. I could be wrong, and please let me know if there is such a trial.
This month, the group led by friend Dhiraj Gupta, in Liverpool, first author Mark Mills, has published an observational study of these devices vs normal care in the Journal of Cardiovascular Electrophysiology, which thankfully is open access.
They used a database called TriNetX that allows access to electronic health records (EHR) in mostly US healthcare organizations.
Like many of these studies, they had two populations — one who had AF ablation with a VCD and the other without.
There were about 14,000 patients in both groups. The primary endpoint was a composite of death, vascular complications, bleeding, and need for blood transfusion within 7 days of the procedure and in extended follow-up.
They did propensity matching to try and match the non-random comparisons.
The primary composite was lower in the VCD group. The relative risk reduction was significant with an odds ratio of 0.76 (CI 0.65 to 0.88). However, the absolute rates were 1.97% vs 2.6%, so that is a delta of 0.63%. or number needed to treat (NNT) of 163.
The difference was driven by lower rates of vascular complications. Again, vascular complications and bleeding all had positively lower relative risks, but in these components of the composite the absolute numbers were even lower. Take bleeding events — 0.9% vs 1.23%. That’s a delta of 0.33% or NNT of 333.
The authors concluded that there is an association between VCD and fewer vascular complications.
Comments. The first thing to say is that the authors avoid causal language, but I side with JAMA’s recent change of heart to allow causal language.
Here, the authors are clearly making a causal effort by comparing these two groups. The question is how well propensity matching of EHRs achieves quasi-randomization. I’d say it’s better than no matching but it’s a long, long way from randomization.
But let’s assume for a moment that this is like an RCT. Let’s take the outcomes as is. Even if the devices are slightly better than non-closure, the absolute risk reductions are tiny relative to costs.
The device costs hundreds of dollars. You need multiple devices per case. We have no direct RCTs, but in this observational study, the control arm rate of vascular complications is approximately 1%, and the reduction is less a half of a percent. You’d have to treat more than 300 patients to prevent one non-fatal groin complication.
An NNT of 300 times a cost of $1000 (2 devices) is $300,000 to prevent a groin complication. This doesn’t make sense. There is no way these could be cost effective.
And we, as a profession, should not be embracing these devices. Yet, I am afraid many centers and operators use them.
SURMOUNT-OSA
I guess I will stay on the cost bandwagon today. As for cost burden to a system, VCDs pale in comparison to GLP-1 agonists.
Last week, the New England Journal of Medicine published a marketing trial called SURMOUNT-OSA.
Tirzepatide vs placebo in adults with moderate-severe obstructive sleep apnea (OSA) and obesity. The primary endpoint is not cardiac outcomes or death, but rather surrogate markers of sleep apnea — namely, the change in apnea-hypopnea index (AHI).
The trial was actually two trials. Trial 1 comprised patients not on continuous positive airway pressure (CPAP) at baseline. Trial 2 comprised patients on CPAP.
The total number of patients was only 469 in the two trials. I remind you here that the cardiovascular (CV) outcomes trial of semaglutide for patients with obesity and established heart disease (SELECT) was more than 17,000.
Well, indeed SURMOUNT did deliver positive results. The AHI was reduced by 20 AHI events per hour in both groups.
As expected, the percent change in body weight favored tirzepatide. Systolic BP was reduced.
Comment. I was going to ignore this trial, but then I thought, well, it is in NEJM. Sleep apnea is an important condition for cardiologists to be aware of. But mostly, I think SURMOUNT provides an important meta-lesson in marketing disguised as science.
The study is supported by Eli Lilly, which has also submitted an application for a tirzepatide indication to treat sleep apnea.
My problem with such a study is that it is a small study with a surrogate marker. Recall that last week, I mentioned David Brown’s paper showing that even myocardial infarction (MI) is an unreliable surrogate. If MI is not a great surrogate marker, change in AHI is surely worse.
What’s more, anything that induces serious weight loss will likely improve AHI. You don’t need AHI to determine that a 16% reduction in body weight is beneficial.
And let me say something else about sleep apnea. Even though it is an important condition, I worry that it is over-diagnosed. While OSA is associated with worse CV outcomes, I struggle to find compelling RCT-level evidence that its treatment reduces actual CV outcomes.
A meta-analysis in JAMA failed to find an effect, and the large SAVE RCT failed to show that CPAP actually reduces CV outcomes. I will also link to a small Norwegian study showing that CPAP reduced AHI but had no effect on AF.
Obviously, severe Pickwickian type disease needs to be treated, but I think we need more RCTs in lesser degrees of sleep apnea.
Surely, the SURMOUNT study is merely a marketing study to try to gain an indication for a common condition. Lilly and Novo-Nordisk have more money than some countries; regulators and clinicians alike should expect them to run properly powered outcomes trials. We know these drugs induce short term weight loss. We need outcomes data over the long term.
© 2024 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Jun 28, 2024 This Week in Cardiology Podcast - Medscape - Jul 01, 2024.
Comments