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 May 31, 2024, John Mandrola, MD, comments on the following news and features stories: The FLOW trial of semaglutide, the DANCAVAS CV screening trial is back, non-invasive cardiac tests (NICT) for chest pain, and conflicts of interest (COI) on social media.
Semaglutide for CKD
Semaglutide has won again. The FLOW trial was a large renal outcomes trial of semaglutide vs placebo in patients who had both chronic kidney disease (CKD) and type 2 diabetes.
Patients with CKD have three proven renal protective agents: RAS inhibitors, SGLT2-inhibitors, and finerenone.
The authors of the FLOW tested yet another one — the GLP-1 agonist semaglutide. Keep in mind that renal outcomes in patients with CKD are much less common than cardiovascular (CV) outcomes, so FLOW also had important CV outcomes results.
Patients had to have CKD. The definitions are quite complex, especially for an old-school cardiologist. I will say them once: CKD was defined by an estimated glomerular filtration rate (eGFR) of 50 to 75 ml/ min/ 1.73 m2 of body surface area and a urinary albumin-to-creatinine ratio [with albumin measured in milligrams and creatinine measured in grams] of > 300 and < 5000, or an eGFR of 25 to < 50 ml/ min/ 1.73 m2 and a urinary albumin-to-creatinine ratio of >100 and < 5000.
Trial procedures included randomization stratified according to SGLT2 inhibitor, which was about 15% in both arms when the trial started. They used an 8-week dose escalation of semaglutide.
The primary outcome was major kidney disease events, a composite of the onset of kidney failure (dialysis, transplantation, or an eGFR of < 15 ml/ min/ 1.73 m2 , at least a 50% reduction in the eGFR from baseline, or death from kidney-related or cardiovascular causes.
Prespecified confirmatory secondary outcomes were tested hierarchically. These included a) total eGFR slope (i.e., the annual rate of change in eGFR from randomization to the end of the trial); b) major CV events (a composite of MI, stroke or CV death), c) death from any cause.
The results. This was a big trial at nearly 400 sites in 28 countries.
The screened-to-enrolled ratio was quite good where more than half screened were enrolled.
There were slightly fewer than 1800 patients in each group. Age 66 years. One-third woman. Average body mass index (BMI) 31.
The mean eGFR was 47. Slightly more than two-thirds were at “very high risk” of kidney disease progression. The mean HbA1c was 7.8. About 95% were on RAS inhibition; 80% were on a lipid-lowering drug.
The trial was terminated early after an interim analysis noted benefits. The median follow-up was 3.4 years.
Semaglutide won. For the primary renal endpoint, it was 5.8% vs 7.5% events per 100 patient years in the semaglutide vs placebo arm. That 1.7% absolute risk reduction (ARR) translates to a 24% relative RR. Hazard ratio (HR) 0.76 (95% confidence interval [CI] 0.66-0.88).
Over 3 years the ARR was 18.7% vs 23.2% so, 4.5%, which translates to a number needed to treat (NNT) of about 20. Quite good.
Lower risk with semaglutide was also observed for a composite of the kidney-specific components of the primary outcome (HR, 0.79; CI, 0.66 to 0.94), as well as for death from CV causes (HR, 0.71; CI, 0.56 to 0.89).
The results were similar in all subgroups, yet the HR was much more impressive in those not on SGLT2 inhibitors vs those on SGLT2 inhibitors (0.73 vs 1.07).
As for the three confirmatory endpoints, they too were all positive. The slope of the decline of the eGFR was better in the semaglutide arm.
The risk of major CV events was 18% lower in the semaglutide arm (HR 0.82 [CI 0.68-0.98]. One peculiar finding was that the rate of CV death drove this benefit.
Myocardial infarction (MI) was lower in semaglutide arm and stroke was slightly higher. It’s curious because those are two major drivers of CV death. And the risk of death was 20% lower in the semaglutide arm (HR 0.80 [CI 0.67-0.95]).
You ask, of course, about weight changes. Indeed, the mean reduction in weight was 4.1 kg more in the semaglutide vs placebo arm. HbA1c was also lower in the semaglutide arm. Systolic blood pressure (BP) was only 2.2 mmHg lower.
Adverse events leading to permanent discontinuation were more common in the semaglutide group than in the placebo group, 13.2% vs. 11.9%. This finding was driven mainly by gastrointestinal (GI) issues. 4.5% vs.1.1%.
Comments: This was a clearly positive trial, stopped early for efficacy. Semaglutide reduced the rate of progression of CKD. It lowered CV outcomes. And it induced weight loss.
Patients with diabetes and CKD surely benefit from semaglutide over placebo.
One question is in which order to use the drugs. RAS-inhibitors, SGLT2-ihibitors, and the non-steroidal mineralocorticoid antagonist finerenone have all shown renal protective properties.
The authors argue that GLP1 agonists should be used early because, “Although studies of SGLT2 inhibitors in patients with chronic kidney disease have clearly identified important benefits with respect to kidney outcomes, the findings regarding effects on major cardiovascular events and death from any cause in this population have been mixed.” (CREDENCE and DAPA-Kidney quite positive for CV death but EMPA-KIDNEY not so much).
I realize GLP1 agonists are costly, but we now have outcomes data in patients with diabetes, in patients with CKD and diabetes, and patients with obesity and established CV disease. And we know the drugs reduce weight — even more than weight loss surgery.
My one caveat and question, and I hope future studies sort this out, is why or how did the FLOW trial show a 2.6% lower rate of CV death with a 0.7% lower rate of MI and a 0.7% higher rate of stroke? What drove the lower rate of CV death? Was this an adjudication issue? I am asking.
Still, for patients with obesity, diabetes and CKD, semaglutide clearly provides statistically robust and clinically important effects in addition to RAS-inhibition.
DANCAVAS at 6 Years
Authors of the DANCAVAS CV screening trial have an interesting post-hoc study published in PLOS-Medicine.
DANCAVAS is about the only screening randomized controlled trial (RCT) that I consider clearly positive. Positive because the primary endpoint was all-cause death — not a disease specific cause of death, like most cancer screening trials.
I will spend time on this trial and post-hoc analysis because I see DANCAVAS as one of the most unique and important CV trials in the last decade. It overflows with learning points.
The main DANCAVAS trial, published in the New England Journal of Medicine (NEJM) in 2022, was a population based RCT of a package of CV screening tests vs standard care. It’s important to point out that the two arms of DANCAVAS were “invited to screening” or standard Danish primary care.
Patients were men between 65 and74 years old. In Denmark, all Danes have a national email. So, the authors were able to randomly assign patients (1:2) to be invited to screening or not.
Patients in the screening group could elect to come for screening or not. They were obviously not blinded. Patients not invited simply did not know they were in a trial.
The screening program was super-interesting. It included a non-contrast CT to look for coronary artery calcium (CAC) and aneurysms; AF screening; ankle-brachial BP; and blood glucose and cholesterol. The primary author, Axel Diederichse,n told me that all this was done usually in a clinic after hours — there were no radiology overreads for the CT scan. It was streamlined screening.
Positive findings were acted on. Lipid lowering therapy and diabetes meds as well as referral for AF or aneurysms were given according to the findings.
Approximately 45,000 men underwent randomization, about 16,000 in the invited group and 29,000 in the control arm.
But. Only two-thirds of those invited actually underwent the screening. To the authors credi,t they analyzed the results according to the 16,000 invited, whether or not they had the screening. It was an intention-to-screen vs no-invitation comparison.
This is important, and it is unlike the NORDiCC colonoscopy screening trial, where they looked at those who were screened.
DANCAVAS authors write in this new paper:
Due to high likelihood of healthy user selection bias, unadjusted comparison of outcomes among those who actually attended the screening and those randomized to usual-care would be highly inappropriate.
The main trial results:
After a median follow-up of 5.6 years, 2106 men (12.6%) in the invited group and 3915 men (13.1%) in the control group had died (HR, 0.95; CI, 0.90 to 1.00; P=0.06).
Stroke was reduced by 7%, MI by 9%, aortic dissection by 5%, and aortic rupture by 19%.
There was a bit of controversy about interpretation of the primary endpoint: NEJM made Dr. Diederichsen write that the trial was negative. But the HR went from 0.90-1.00 and the P-value was 0.06. So, if you went with the PROTECTED-TAVR trial interpretation, the conclusion would have included the phrase, “On the basis of the 95% CI, the results may not rule out a benefit of the CV screening program.”
Now to the post-hoc paper. Here they looked at 6 years of outcome data (it was 5 years in the main paper). There was one other novel twist to this post-hoc analysis: the authors decided on a primary endpoint by asking patients what they valued most. They sent out a survey to a random group of about 9,000 patients.
The two highest ranking outcomes from the patient survey was MI and stroke. So, that was the primary endpoint for this analysis.
Efficacy: (MI/stroke primary endpoint)
10.8% in the invited-to-screening group vs 11.5% in the usual-care group experienced MI or stroke (HR 0.93 (CI 0.88 to 0.98; P = 0.010). That is, the number needed to be invited to screening was high at 148.
They also included a secondary composite major adverse cardiac event (MACE) of stroke, MI, HF, CV death.This was positive as well for screening: 6% reduction; HR 0.94 (CI 0.89 to .98 P=0.009
They included another composite called MALE (major adverse limb events; aortic dissection or rupture, revascularization due to limb ischemia, amputation due to peripheral artery disease or CV death). This was 9% reduced with screening; HR 0.91 (CI 0.84 to 0.99).
Initiation of antiplatelet and lipid-lowering agents was increased in the invited-to-screening group, while there were no differences in initiation of anticoagulants, hypertensive and antidiabetic agents.
Elective aneurysm repair was more common in the invited-to-screening groups (0.9% vs 0.6%; HR, 1.58; 95% CI, 1.26 to 1.97).
Rather than looking directly at those screened (per protocol analysis), they did an inverse probability of treatment weighting (IPTW) analysis to sort out those most likely to accept screening in the two groups. This is hard for me to totally understand but I think it’s a way of matching those in the control arm vs those in the active arm who would have had screening. The propensity score p(X) is the conditional probability of attending screening given pre-screening characteristics.
This also yielded a 13% statistically significant reduction in the primary endpoint of stroke and MI for screening.
Safety
For intracranial bleeding, it was 1.9% in the screening group vs 1.7% usual care. That tiny difference however calculated to a 17% higher relative risk increase and it was statistically significant (HR 1.17; 1.02-1.35) p= 0.03)
Severe GI bleeding was not different statistically, 5.9% vs 5.8%.
Comments on the 6-year results. I wrote about the main trial in 2022. I noted how unique this trial was. First, this was not CAC screening. It was a comprehensive package of very efficient screening, done after normal clinic hours. A quick CT scan was done only for measurements by a technician; during the scan a rhythm strip checked for AF. A blood draw and a BP were taken and the patient was home in short order.
It’s Denmark, so there’s no culture of coronary angiography and stress testing in asymptomatic people and as such there was no increase in percutaneous coronary interventions (PCI) or coronary artery bypass grafting (CABG).
The take-home for the 6-year results was that if you do such a screening in an environment like Denmark, you can reduce CV outcomes.
The reason for the reduction in CV events may be the higher use of antiplatelets and lipid-lowering drugs, or it may be behavior change — behavior change from discovering abnormalities or meeting a kind and caring person such as Axel Diederichsen.
It doesn’t matter why there were better outcomes. That’s the beauty of trials.
In the paper the authors make note of a higher rate of intracranial bleeds. On the surface this seems like a counter-force against the benefit, but it’s not super-relevant because the rate of intracranial bleeding is — fortunately — quite low.
The rate of MACE outcomes was 30%, the rate of intracranial bleeding was less than 2%. And the delta was only 0.2%.
The journal PLOS Medicine has the authors include a major limitation in the abstract, which I like, and here it was the fact that DANCAVAS was restricted to men between the ages of 65 to 74. The results therefore do not generalize to women or men of other age groups.
I’ve discussed this in my 2022 column. The reason DANCAVAS authors chose this select patient group is that this age group of men had the highest rate of CV events and lowest rate of competing risks. Women, they argue, would have had such a low rate of CV events that screening would have had little chance to reduce outcomes. They even did a pilot trial showing this.
That enrollment decision provides a shining example of how most trials select highly specific groups of patients. This is not nefarious. It’s common sense.
It’s why you don’t see lots of frail elderly patients with end stage renal disease in most cardiac trials — because these patients have too many competing risks of outcomes.
It’s also why cardiac trials enrich their patient population with patients likely to have a cardiac event. The ODYSSEY OUTCOMES PCSK9i trial of alirocumab included patients who were post-acute coronary syndrome (ACS), and the SELECT trial of semaglutide included patients with obesity and established CV disease.
The point is that trials need to include patients who are at risk of having the primary endpoint that will be studied.
Our job as users of the evidence is to smartly consider who was in a trial, what was the trial environment like, and how clinically important were the results. DANCAVAS requires smart thinking for translation.
Here I use my science hat and US clinician hat.
Science wise, DANCAVAS is also notable because background care in Denmark is good. So, the highly unique screening program won against a quite good comparator.
But using my US-clinician hat: These results do not translate to the United States wherein all CT scans are over-read by a radiologist, which would lead to a flurry of incidental-omas.
Also, US CAC scans often lead to stress-testing, angiogram, PCI, and CABG cascades — another drain on efficacy and cost efficacy. It’s a drain on efficacy because there are numerous trials showing no benefit for revascularization in asymptomatic coronary artery disease (CAD) over medical therapy.
Finally, I absolutely love the take of DANCAVAS authors in strongly stating how wrong it would be to look at only those who received screening, because clearly, these are different types of patients. This was not done in the NORDICC trial and it allowed the GI associations to provide talking points that screening works if you get it done (but even then, there was no difference in overall death).
Stress Testing for Chest Pain
This paper from authors in the Kaiser Permanente Network looked at emergency department (ED) referral for non-invasive testing patterns in patients with chest pain. The first author was Dr. Dustin Mark.
It’s an observational study but the authors looked to associate 2-year hard clinical outcomes (MACE outcomes) and referral patterns. MACE included MI, cardiac arrest, shock, or CV death.
It’s an important question because a) chest pain is a common a presentation, b) you don’t want to miss a life-threatening coronary lesion, c) numerous trials show no benefit to revascularization in stable CAD, d) non-invasive testing often picks up incidental CAD unrelated to the chest pain, and e) non-invasive testing is costly.
Circulation Outcomes published the retrospective cohort study conducted in 21 EDs over 6 years.
These were non-MI patients. Patients with ST elevation MI and positive troponin chest pain were excluded.
They grouped patient encounters by the NICT referral rate of the initially assigned ED doctor relative to local peers. There were low, intermediate, or high referrers. They also looked at how the heart score of patients correlated with outcomes.
This was a large study including144,000 58-year-old patients with chest pain. More than half were women.
The rate of NICT referral at 30 days was 13%, 20%, and 28% in the low, intermediate, and high NICT referral groups respectively.
Patients in the three non-randomized groups were well-balanced on baseline characteristics.
Compared with the low NICT referral group, there was no significant decrease in the adjusted HR of MACE within the intermediate (adjusted HR, 1.08 [CI, 1.02 to 1.14]) or high (adjusted HR, 1.05 [CI, 0.99 to1.11]) NICT referral groups.
Some criticized this analysis because it was simply looking at referral patterns of doctors. But the authors write that actual NICT were higher, 53% vs 33% in the high vs low referral arm.
In addition, they also analyzed outcomes according to HEART score which puts patients into a low, moderate, or high-risk category. Here there also were no differences in outcome among the low vs high referral patterns.
Comments. First, not only was there not an association for benefit, but there was also actually a trend toward harm from being an intermediate or high referrer, though in adjusted analyses this did not reach statistical significance. And it is observational and I would not suggest harm.
But I do like this paper. It’s a bit of a quasi-randomized study because who a doctor sees — a low, intermediate, or high referring ED doc — is a bit random. And there was not an association of better outcomes for those who referred more, even when considering higher HEART score patients. And baseline characteristics were matched.
This lack of association makes sense to me, because think about chest pain without troponin elevations in the modern era. Troponin assays are so sensitive that if there is no elevation it is almost certain that the chest pain is not heralding an unstable ACS — which is the point of risk stratification with any means — be it functional with NICT or anatomic with CT angiography.
The vast majority of NICT were stress ECG or perfusion scans.
Referral patterns not associating with better outcomes also make sense because any CAD picked up will surely be incidental (given the negative troponins) and if it is incidental, it is asymptomatic and revascularization would have no benefit over medical therapy.
I realize the SCOT-HEART proponents will say that picking up disease — even if incidental — is beneficial because diagnosis of disease will cause initiation of medical therapy, especially since these were young 58-year-olds. Indeed, that was the argument as to why CT angiography scans in SCOT-HEART were found to be beneficial, though I am not sure.
In this analysis, slightly more medical therapy was started. That could be a positive, but the SCOT-HEART benefits did not replicate here.
But of course, I, like the authors, have to acknowledge the limitations of this non-random analysis. We should be just as skeptical of observational studies that find no association as we do the ones that find positive associations.
That said, I think we do way too many coronary evaluations in troponin-negative patients with chest pain. This would be a great area for a Patient-Centered Outcomes Research Institute (PCORI)-type pragmatic study randomizing high-sensitivity troponin negative patients to NICT or no NICT.
COI and Social Media
Yesterday, JAMA published a research letter on industry payments to doctors who endorse drugs and devices on social media.
It’s a small study, led by a Sloan – Kettering team; first author Sonia Persaud.
As background, their first sentence opens your eyes: “Physicians commonly receive financial payments from industry, totaling $2.46 billion in 2022.”
They used drugs approved in 2021 and 2022. They basically linked each doctor to Open Payments. The social media platform studied was Twitter or X.
They included posts from physician and industry accounts, whether the promoted product was explicitly named or implied.
Of 28 physician endorsers in 2022, 75% specialized in urology, medical oncology, or radiation oncology.
The median H-index (a measure of an author’s publication productivity and impact) was 15 (interquartile range, 3 to 33); 68% (19 of 28) had an H-index less than 20, so these are mostly non-academic doctors.
Half the endorsers had no publications related to the endorsed drug or device.
Two-thirds were sponsored physician testimonials, with half having no disclosure of compensation.
The specialty-weighted mean of general payments received in 2022 across all physicians was $4500.
More than 90% received payments from the endorsed products company.
Comments. Take a look at the letter. Good on JAMA for publishing this. I was part of a recent letter documenting high percentages of docs who receive money from industry as well as a serious skew, in that some receive lots of money from industry.
Social media is another new source of advertising. It’s especially useful for industry to enroll doctors as spokespeople on social media.
Though this study did not focus on cardiology, I mention it in hopes that it creates awareness of a trend that I think is unseemly and borderline unprofessional. That is, the trend in cardiology for industry to take pictures with us and post online that we did our first, or our 100th or 500th procedure with the company’s product.
The picture includes the shiny happy industry people and doctors who stand in front of a company logo.
I am not against industry collaboration, but we should be mindful of being used by industry. And social media is another easy way that industry advertises, sometimes, often even, products with dubious benefits or cost efficacy.
© 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: May 31, 2024 This Week in Cardiology Podcast - Medscape - May 31, 2024.
Comments