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 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 August 12, 2022, John Mandrola, MD comments on the following news and features stories.
Using Wearable Devices to Detect AF Is 'Cost Effective'
The Journal of the American Medical Association (JAMA) Health Forum published a study on the cost-effectiveness of screening for atrial fibrillation (AF) using wearable devices. The same week, The New England Journal of Medicine (NEJM) published a ‘clinical decisions’ article centered on the question of whether to do AF screening in a 75-year-old man with hypertension and diabetes. One extremely accomplished professor argued for AF screening. One marginal podcast host argued against further AF screening.
Let’s talk about AF screening because it’s coming up almost every week in clinic — and not just in the electrophysiology (EP) clinics, but also in general cardiology clinics, emergency departments, and primary care clinics.
The general idea of screening for disease comes up often on #TWICPodcast but these have always been about medically directed screening — stuff we do: electrocardiogram (ECG) monitors, colonoscopies, mammography, vascular screening etc. The unique thing that’s happening in AF screening now is that doctors are being shut out of the picture because of consumer devices, like smart watches. I am seeing more and more patients with screen-detected AF that came not from a medical-grade monitor, but a consumer device. The article in JAMA Health Forum was an economic evaluation using simulation methods of different strategies.
They considered eight AF screening strategies; six of the eight used consumer devices and two used traditional methods, such as pulse palpation or 12 lead ECGs vs no screening.
They then modeled multiple scenarios based on age to start screening, male vs female, daily wear time and cost of the device.
The primary outcome measure was the incremental cost effectiveness ratio (ICER). Ratios being fractions this was US dollars divided by quality-adjusted life years (QALY).
Secondary measures were stroke and bleeding.
They report that all six screening strategies using wrist worn devices were more effective than no screening.
The screening model with the lowest QALY gain was the simple 12 lead ECG.
The screening model with the best QALY gain was a sequence of a wrist-worn wearable pulse monitor followed conditionally by wrist-worn wearable one-lead ECG and confirmatory patch monitor.
Strategies using wrist-worn wearable devices were associated with reduction in stroke incidence by 20 to 23 stroke events per 100,000 person-years but an increase in major bleeding by 20 to 44 major bleeding events per 100,000 person-years. Note the small numbers – 23 strokes and 44 bleeds per 100,000 person-years.
The overall most cost-effective strategy was wrist-worn wearable photoplethysmography (PPG) followed conditionally by wrist-worn wearable one-lead ECG and confirmatory patch monitor (ICER, $57,894 per QALY).
The ICER of the most cost-effective strategy increased as the minimum age of the cohort decreased starting with individuals age 60 years or older This is not surprising because you are more likely to find AF in older people. Also not surprising was that the ICER of the most cost-effective strategy increased as the cost of the wearable increased. Cheaper devices did better.
The authors conclude that “This decision-analytic model suggests that screening using wearable devices is cost-effective compared with either no screening or AF screening using traditional methods.”
Comments. The authors were extremely candid about their limitations. It’s a simulation economic analysis. Models have to make assumptions. Their limitations paragraph was laudable. Do read it. My problem is higher up than any specifics. Here, the assumptions are existentially flawed: namely, how do you do a cost efficacy analysis if there is not a shred of data that there is efficacy?
The adequately powered LOOP trial, which used a far better AF detection device than a consumer wrist device, an always-on implantable loop recorder, failed to find a statistically significant reduction in stroke; no reduction in deaths; and a non-significant 26% increase in major bleeding.
You might ask why I cover an economic modeling paper? Because I think we need to have a laser-like focus of any empirical assessment of consumer devices—which are shockingly devoid of outcomes data. The question is not whether AF will be detected, it surely will. The question for patients and doctors alike is how best to handle this new information and whether these devices improve the health of the consumers or of the manufacturers.
I don’t exactly know, but I can say this, at least about our system: the more you come to know about all the uncertainties surrounding AF, the more frightening it is to think about the millions of asymptomatic people who will now be exposed to a profit-driven healthcare system, where many if not most clinicians are infected with action bias.
The stress tests that lead to angiography and then stents;
The echocardiograms that lead to all-manner of workups for shadows;
The potential excess use of anticoagulation;
The extra AF ablations.
All the above without any data — except for LOOP and STROKESTOP, which strongly suggest that medical-grade screening is ineffective in improving health.
My advice to nine out of ten smartwatch owners is to give it away. I don’t say this to patients, but I will leave this topic with a quote from the late Dr. Petr Skrabenek: “The pursuit of health is a symptom of unhealth.”
Sodium, Potassium, Salt Substitutes and CV outcomes
More Evidence Salt Substitutes Lower Risk of CVD and Death
The journal Heart, a subsidiary journal of the British Medical Journal, has published a meta-analysis of trials that looked to assess the effects of salt substitutes on multiple outcomes, not just blood pressure (BP), but cardiovascular (CV) outcomes. Sodium, potassium, and blood pressure are obviously vital. But this is also a super-interesting topic scientifically because salt substitutes do two things: they reduce sodium, but they also increase potassium. Sorting the effects of the two may be very important in informing dietary advice. Yet there is even more to this meta-analysis: that is the question of how generalizable is the massive, 20,000-patient-strong SSaSS cluster randomized clinical trial published in NEJM last September. SSaSS was performed in rural China; did it apply to other countries?
Before I tell you about the meta-analysis, recall that SSaSS randomly assigned 600 villages in China to a 75/25 sodium-potassium salt substitute or to a control arm.
Participants in the trial were high risk, as they had to have a history of stroke, older than 60, and with hypertension (HTN).
The primary outcome was stroke. Period. Secondary outcome was major adverse coronary events (MACE). SSaSS was hugely positive:
The intervention arm had a statistically significant 14% reduction in stroke. The P-value was very low.
MACE and death were also reduced significantly.
However, systolic BP reductions were modest at only 3 mmHg.
This study created buzz amongst the lower-sodium-proponents crowd. Trial discussant Dr. Bryan Williams said this at the European Society of Cardiology (ESC) meeting when it was presented:
“Those who doubted the benefits of a salt restriction for cardiovascular disease prevention were wrong, the debate stops here.”
Now, I am not so sure, and let me use the Heart meta-analysis and an editorial in the European Heart Journal to explain.
The authors of the meta-analysis found 21 trials that reported the effects of salt substitution on BP or clinical outcomes. Nineteen trials reported on BP and five reported clinical outcomes.
The total number of participants was 32,000. (One of the trials was SSaSS and that had 20,000 patients.)
Overall reduction of systolic BP was −4.61 mmHg.
There were also clear protective effects of salt substitutes on clinical outcomes as well. Reductions in total mortality, CV mortality, and MACE all were statistically significant.
Reductions in BP were consistent across geographical regions and population subgroups no matter the age, sex, history of hypertension, body mass index, baseline BP, baseline 24-hour urinary sodium, and baseline 24-hour urinary potassium.
The authors did a meta-regression, which is basically a correlation that attempts to account for the effects of different variables. The meta-regression found that each 10% lower proportion of sodium chloride in the salt substitute was associated with incrementally greater reductions in BP.
But here’s the thing: nearly all the salt substitutes had a sodium:potassium ratio. As you went down in sodium, the general trend was to go higher in potassium.
And in supplement Figure 5, the plot of studies looking at urinary sodium and potassium finds that nearly every study but one finds a significant increase in potassium excretion, while most of the studies find no significant decrease in sodium excretion. It really makes you wonder whether it’s the lower sodium or boost in potassium that is doing it.
Now I want to point you a viewpoint published this month in the European Heart Journal, first author, Prof Franz Messerli, and senior author, Salim Yusef.
They make the provocative but persuasive case that the major reductions in outcomes seen in SSaSS and this meta-analysis stem mostly from potassium increases, and that efforts to highly restrict daily sodium intake may be misguided.
They analyzed the effects on salt substitution in SSaSS on 24-hour sodium and potassium excretion. They then compared those to the values of the average level of intake in the United States as determined from published National Health and Nutrition examination survey (NHANES) data.
At baseline in SSaSS the average sodium intake was very high at 4.3 gm per day. The potassium intake was very low at 1.4 gm per day.(Video) ESC 2022: The EMMY Trial: Empagliflozin in Patients with Acute MI | Prof Harald Sourji
This is higher in sodium (3.6 gm) and lower in potassium (2.1 gm) than the average American intake.
In the active arm of SSaSS, the salt substitute arm, sodium was reduced by only a modest 8% from 4.3 to 3.9 gm. That’s hardly anything and 3.9 gm is still higher than the average American intake of 3.6 gm.
That’s also double the recommended intake by WHO and ESC standards and three times what the American Heart Association recommends.
Now for the punchline:
The SSaSS intervention, the salt substitute, increased potassium excretion by a giant 57%, up from 1.4 gm to 2.2 gm, which gets them to the mean American intake of potassium.
They then point out that an 8% reduction in sodium would only reduce systolic BP by 1 mmHg while a 57% increase in potassium would be expected to reduce systolic BP by 3 mmHg.
This makes sense because a potassium deficient diet would increase sodium reabsorption in the kidney, further exacerbating excess sodium intake.
“Therefore,” they conclude, “It is highly likely that most, if not all benefits observed in SSaSS were due to the increase in potassium with the salt substitute. Importantly, the dietary Na+/K+ ratio decreased from 3.1 to 1.8 with the salt substitute intervention.”
The policy implications they get to is that SSaSS does not inform the controversy regarding the guideline recommendations for low-sodium diets. And surely there is no support for a 2 gm per day sodium intake. And given the lack of hyperkalemia signal in SSaSS or the recent meta-analysis, they find reassurance in the efficacy and safety of increasing the potassium intake.
Finally, a meta-lesson here on critical appraisal. I interviewed SSaSS principal investigator Bruce Neal when the trial came out and I did not press him on the absolute values of sodium and potassium intakes. Messerli and colleagues show the importance of looking at what actually happens in a trial. Yes, the SSaSS intervention reduced sodium intake, but the absolute value was still super high. The bigger difference was in the absolute changes in potassium.
Two examples of how details matter to interpretation of trials.
The REDUCE-IT trial of fish oil vs placebo in patients with high triglycerides. Yes, the fish oil group reduced outcomes relative to the placebo group, but there were only modest reductions in triglycerides.
But in the placebo group, there were increases in lipid and inflammatory markers suggesting, strongly, that at least a lot of the effect was due to placebo harm rather than active treatment benefit.
The original FAME trial found that fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) reduced outcomes relative to angiography-guided PCI. This launched FFR as an approach.
But was it that FFR was special? Or was it the fact that far-fewer PCIs were performed in the FFR group? One way to interpret FAME is that FFR is great. Another way is that more PCI led to harm.
Given that these were patients with stable coronary artery disease, I find it highly likely that doing less PCI was the driver of benefit.
House Could Soon Give Biden Win on Medicare Drug Negotiations, Insulin Costs
Something super-rare may happen this month. Sit down for this: the US Congress may get together and actually do something to improve American healthcare.
I can’t believe it, but it’s possible that a bill that allows Medicare to negotiate drug prices may pass and it may lead to lower drug prices, if not directly, also through indirect pressure on the drug makers.
I am nowhere near qualified to get into the policy weeds on this, but I know that in cardiology, affordability of evidence-based drugs is a serious problem in this country. Yes, I often rail about low-value care, and there is a lot of it out there — things like thrombolytics for acute stroke and icosapent ethyl for hypertriglyceridemia. But basic oral anticoagulants for AF, sacubitril/valsartan and SGLT2 inhibitors for patients with heart failure with reduced ejection fraction are often unaffordable to those who could most benefit. So allow me to be slightly optimistic about government doing something to help. Emphasis here on slightly optimistic.
Fitness Is Underrated
Cardiorespiratory Fitness Key to Longevity for All?
Let’s close by talking about cardiorespiratory fitness (CRF). First, my bias: I am addicted to endurance exercise. Mostly cycling at the moment, but also running. Even swimming in past years. That said, I strongly suspect that exercise as an intervention gets short shrift among doctors. Two recent studies deserve brief comment.
JACC has published a neat study using Veterans Administration (VA) data. They looked at older veterans who had treadmill exercise tests. They excluded people with heart disease. VA data allows for using METS achieved as a marker of CRF and death.
For every category of age—even septuagenarians and octogenarian s— there was an inverse and graded association between fitness and death.
Some fitness was better than low fitness. Moderate fitness was better than low. High fitness was better than moderate.
The super-fit showed no association with higher death rate.
This was an observational study, and those with high CRF also do other things to extend life, like eating a heathy diet. Also, there is real chance of reverse causality, meaning that the luck to live longer is also why you do better on treadmill tests.
Before I say anything, else let me tell you about another CRF study, from the Adelaide group of Prash Sanders.
Patients with paroxysmal AF (PAF) or persistent AF had exercise testing and echocardiography. Included were patients in sinus rhythm with normal ejection fraction (EF) who were able to complete the stress test.
Investigators measured all manner of left atrial (LA) function from echo. CRF had strong associations with better LA function; less LA stiffness for instance. No this paper does not establish cause and effect, but it is another step in understanding why normal exercisers have far less AF.
Exercise data will surely be heavily observational. Yet I still believe there is enough signal to prescribe exercise as a prevention therapy or an anti-arrhythmic or anti-hypertensive.
I tell patients that they should exercise every day that they eat, for 15 to 30 minutes. Consider it like a pill.
Compliance is low, but when patients do it, the effects are amazing. Also – imagine a world where every clinician recommended it. Norms might change.
Though we don’t have randomized controlled trials, I bet this recommendation stands up to time better than most things we do in cardiology now.