atrial fibrillation: treating subclinical AF, new AF guidelines

 A recent article found that short-lasting, asymptomatic intervals of atrial fibrillation (“subclinical atrial fibrillation”) were associated with a significantly increased risk of stroke; and apixaban, vs aspirin, was effective in decreasing stroke or systemic embolism, though at a cost of increased bleeding (see afib subclinical apix helps NEJM2023 in dropbox, or DOI: 10.1056/NEJMoa2310234)

 

Details:

-- 4012 patients from 247 clinical sites in 16 European and North American countries who had subclinical atrial fibrillation (SCAF) lasting 6 minutes to 24 hours, as determined from their long-term continuous monitoring with pacemakers or defibrillators, were randomized to apixaban 5 mg twice a day (2.5 mg if indicated) versus aspirin 81 mg daily, in the ARTESIA trial

-- eligible participants also a CHA₂DS₂-VASc score of at least 3, and subsequent protocol amendments raised the minimum age to be at least 55yo and allowed for the enrollment of patients who were at least 75yo who had had a stroke with or without other risk factors

-- the trial medication was stopped and anticoagulation started if the SCAF became clinical or lasted more than 24 hours

-- mean age 77, 36% women, 94% white European, weight 83 kg

-- blood pressure 135/76 mmHg

-- pacemaker 69%, ICD 14%, CRT-ICD or CRT-pacemaker 12%

-- CHA₂DS₂-VASc score 3.9 (score ranges from 0 to 9 with higher scores indicating higher risk of AF)

--  history of  major bleeding >6 months before enrollment: 2%

-- baseline medical comorbidities: hypertension 82%, history of coronary artery disease 37%, peripheral arterial disease 8%, diabetes 29%, heart failure 28%, history of stroke or systemic embolism or TIA 9%

-- medication use: aspirin 57%, other single antiplatelet agent 4%, dual antiplatelet therapy 3%

 

-- Number of episodes of SCAF in prior six months: 0 in 18%, 1-5 in 64%, 6-50 in 17%, > 50 in 4%

-- longest period of SCAF in prior six months: 6  min to 1 hr in 26%, 1 to <6 hours in 36%, 6 to <12 hours in 14%, 12 to 24 hours 7%

-- Primary efficacy outcome: stroke or systemic embolism

-- additional outcomes included cause-specific mortality; stroke subtype; and TIA with motor deficits, aphasia or duration of more than 5 minutes

-- primary safety outcome: major bleeding in the on-treatment population (eg, all of the patients who underwent randomization, including those who dropped out of the study because of developing atrial fibrillation, AF, for at least 24 hours or symptomatic AF)

-- mean follow-up 3.5 years

 

Results:

-- over 18.3 months of followup, 24% of the participants developed subclinical AF lasting >24hrs or clinical AF, leading to their stopping the trial meds

-- death occurred in 22% of participants

-- 9.4% of the patients were on the lower dose of apixaban

 

-- stroke or systemic embolism (intention-to-treat analysis, which included those who were censored from the full study because of progression to clinical AF or >24 hours of subclinical AF):

    -- apixaban group: 55 events, 0.78% per patient-year

    -- aspirin group: 86 events, 1.24% per patient-year

        -- 37% decrease with apixaban, HR 0.63 (0.45-0.88), p=0.007

-- ischemic stroke and stroke of unknown cause:

    – 38% decrease, HR 0.62 (0.43-0.91)

-- stroke from any cause:

    – 36% decrease, HR 0.64 (0.46-0.90)

-- hemorrhagic stroke:

    –  24% decrease (though only 10 in apixaban group and 13 in aspirin), HR 0.76 (0.33-1.73), not statistically significant

– stroke severity (using the Rankin scale, where a score of 3 to 6 included disabling to fatal strokes):

    –  18 of 55 (33%) on apixaban vs 36 of 84 (43%) on aspirin

 

-- stroke or systemic embolism (on-treatment patients who continued for the full study):

    -- apixaban group: 36 events, 0.71% per patient-year

    -- aspirin group: 65 events, 1.29% per patient year

        -- 45% decrease, HR 0.55 (0.37-0.83), p=0.004

-- ischemic stroke and stroke of unknown cause:

    – 46% decrease, HR 0.54 (0.35-0.86)

-- stroke from any cause:

    – 43% decrease, HR 0.57 (0.38-0.85)

-- hemorrhagic stroke:

    –  22% decrease (though only 8 in apixaban group and 10 in aspirin), HR 0.78 (0.31-1.98), not statistically significant

– stroke severity (using the Rankin scale, where 3 to 6 is for disabling to fatal strokes):

    –  11 patients on apixaban vs 29 on aspirin, 63% decrease, HR 0.37 (0.19-0.75)

 

-- major bleeds:

    -- apixaban group: 1.71% per patient year

    -- aspirin group: 0.94% per patient year

        -- 80% increase, HR 1.80 (1.26-2.57), p=0.001

    – of note: only 17 episodes in patients on apixaban vs 13 on aspirin; and 9 of the patients on apixaban required immediate measures to avoid death or died from bleeding vs 4 on aspirin

-- fatal bleeds in those completing the study:

    -- apixaban group: 5 patients

    -- aspirin group: 8 patients

-- symptomatic intracranial hemorrhage in those completing the study:

    – apixaban group: 12 patients

    – aspirin group: 15 patients

 

Commentary:

-- this study was based on pacemakers and implantable cardioverter-defibrillators since they have been around long enough to have long-term data for lots of patients, and not the more recent implantable loop recorders etc

-- the authors of the current study had done a prior study on 2580 patients who were at least 65yo, had hypertension without AF, and had a recent pacemaker or defibrillator; they were monitored for 3 months, finding that subclinical atrial tachyarrhythmias (>190 beasts/minute for more than 6 minutes) had a more than 5-fold risk of developing clinical AF (HR 5.56),  and >2-fold risk of ischemic stroke or systemic embolism (HR 2.49); the population attributable risk of stroke or systemic embolism associated with subclinical atrial tachyarrhythmias was 13%: https://www.nejm.org/doi/full/10.1056/nejmoa1105575

    – a study of 590 people from the general population who had stroke risk factors ( at least 70yo, and at least one of hypertension, diabetes, heart failure or prior stroke) but without AF had >3 years of continuous loop recording (659,758 days of data) discovered 20,110 episodes of AF lasting at least 6 minutes. These recorded episodes were found especially in those at least 75yo, males, CHADS₂ score >2, or NT-proBNP >40pmol/L. these were recorded AF episodes lasting at least 24 hours vs <24h. this study found that loop recorders were much better than ECG monitoring performed for 24-hour or 48-hour or 7-day or 30-day continuous monitors. There was improved sensitivity for AF detection by longer screening and higher frequency of repeated screenings up to annually. The absolute increase in stroke risk was half that of those with clinical AF: https://www.ahajournals.org/doi/epub/10.1161/CIRCULATIONAHA.119.044407

    -- another more recent study (STROKE AF: see stroke cryptogenic AF jamaneuro2023 in dropbox, or doi:10.1001/jamaneurol.2023.3931) assessed  patients with prior ischemic stroke in 33 sites in the US who had either an insertable cardiac monitor (ICM, eg loop recorders) versus usual care, confirmed an increasing detection of atrial fibrillation over a three-year period in 21.7% in the ICM group versus 2.4% in the control group, a hazard ratio of 10.0. Of note the ICM group had increased detection of AF progressively over that three-year period, with the highest AF burden of 33.9% in the participants who had AF durations of six minutes per day, and 22% in those with AF burden of 30 minutes/d (“usual care” was defined as site-specific usual care, without further elaboration)

 

-- the new atrial fibrillation guidelines were just released with a huge shift from just treating already identified atrial fibrillation to understanding that AF is a long-standing process with a progression through stages.  their recommendation is to treat modifiable risk factors independent of the stage (see atrial fibrillation guidelines circ2023 in dropbox or DOI: 10.1161/CIR.0000000000001193) 

    -- first stage is “at risk for AF”:

        -- Non-correctable risk factors include advancing age, family history, and being taller.  The correctable risk factors (with mendelian randomization studies supporting them as risk factors) include smoking, sedentary lifestyle, alcohol (dose-response curve beginning at 1 drink/day, having a relative risk of 1.08 and increasing with higher alcohol consumed), adiposity (RR 1.28 for each 5 unit increase in BMI), hypertension  or hypertension treatment  (studies suggest that SBP<120 decreases risk, as does use of mineralocorticoid receptor antagonists), resting heart rate (J-shaped curve with optimal being 68-80 bpm) and diabetes (increased risk of AF, with suggestion that optimizing glycemic control prior to ablation, their suggested first-line therapy in selected patients, decreases AF recurrence subsequently)

    --second stage is “pre-AF”, suggesting heightened surveillance:

        -- cardiovascular: heart failure (bidirectional relationship), coronary artery disease, valvular heart disease, cardiac surgery

        -- other conditions: chronic kidney disease, obstructive sleep apnea, hyperthyroidism, sepsis

        -- markers on ECG: prolonged PR interval, LVH

        -- biomarkers: natriuretic peptides, inflammatory markers (CRP, IL-6, TNF-a), Lp(a)

        -- imaging markers: LA size and function, LV wall thickness

        -- social determinants of health: education (Danish study finding decreased risk in higher educated young individuals), income (Danish study finding decreased risk in those with higher income), low SES

            -- of note, the correctable risk factors above as well as the social determinants of health ones here are basically all associated with chronic inflammation, which is associated with a notable increase in cardiovascular disease development (see https://gmodestmedblogs.blogspot.com/2023/10/update-ascvd-risk-factor-critique.html )

    -- third stage is the progression of AF that patients may have found by clinically appropriate monitoring for AF burden

        -- paroxysmal AF (intermittent and terminates within 7 days of onset)

        -- persistent AF (continuous AF at least 7 days and requires intervention)

        -- long-standing AF (AF that is continuous for at last 12 months)

        -- successful AF ablation

            -- all of these stages should involve ongoing monitoring as clinically appropriate for AF burden,  determining if the AF is associated with pathophysiological changes, performing stroke risk assessment and appropriate therapy, and treating symptoms:

        -- fourth stage: permanent AF, with no further attempts for rhythm control after discussion between patient and clinician

    -- clinical evaluation:

        -- risk stratification/screening: there are >20 risk prediction models for developing AF (the CHARGE-AF one being the most widely replicated, which includes age, white race, height, weight, systolic and diastolic blood pressures, smoking, diabetes, and history of MI)

        -- important tests include the usual culprits (EKG, thyroid test, electrolytes, EKG, echocardiogram), and detecting AF for diagnosis as well as AF duration by the  higher sensitivity tests with longer monitoring (through ambulatory electrocardiographic monitoring and implantable loop recorders; as noted above the most sensitive evaluation seems to be loop recorders for 3 years, though there have not been any studies validating differences in outcomes as a result of this long-term assessment)

            -- smart watch apps only rarely detect new onset AF and should not be relied upon

    -- the guidelines also recommend early rhythm control to maintain sinus rhythm and minimizing AF burden

 

-- this ARTESIA study above found that apixaban was superior to aspirin in those with subclinical AF for decreasing both the number of strokes and their severity (disabling or fatal). And though there were more major bleeds with apixaban, there actually were fewer really severe ones (fewer fatal or intracranial bleeds) but more fixable ones (GI bleeds, need for transfusion)

-- though it is certainly important to assess both the positive outcomes (benefits) and the negative ones (risks), it is also important to put these in context: the impressive apixaban benefit translated into many fewer strokes and, importantly, an even lower percent of severely disabling or fatal strokes; and the serious adverse events were predominantly GI bleeds and the necessity for transfusions, both of which were fixable (the apixaban group, though having more "major bleeding" actually had a total of 17 really severe bleeds (fatal and symptomatic intracranial hemorrhage) vs 23 on aspirin

 

--  another similar trial (NOAH-AFNET) was also recently published on patients with atrial high-rate episodes, but coming to the opposite conclusion: see  https://www.nejm.org/doi/full/10.1056/NEJMoa2303062

    -- 2536 patients at least 65yo, with atrial high-rate episodes (AHREs) detected by implanted cardiac devices and lasting at least 6 minutes who also had at least one additional risk factor for a stroke, were randomized to edoxaban 60mg (reduced to 30mg if <60kg, or eGFR 15-50, or were on were on strong P-glycoprotein inhibitors;  this reduction was done in 28.7%) vs placebo

    -- primary efficacy outcome was composite of cardiovascular death, stroke, or systemic embolism; safety outcome was death from any cause or major bleeding

    -- study stopped at 21 months because of safety concerns:

        -- primary efficacy outcome:  3.2% per patient-year on med vs 4.0% in placebo group, 19% better with med, HR 0.81 (0.60-1.05), a not statistically-significant trend to benefit

        -- primary safety outcome: 5.9% per-patient year on med vs 4.5% in placebo group: 31% increase, HR 1.31 (1.02-1.67, p=0.03)

    -- but there are several concerns about the quality of this test:

        -- the major bad outcome was in “death from any cause” in 111 vs 94 patients, with the major specific contributors being cardiovascular death (52 on med vs 57 on placebo), followed by cancer-related death (22 on med and 9 on placebo)

            -- but, these people were at quite high risk of a cardiac event anyway, given their age and risk factors, so from a statistical perspective, the numbers of adverse events found seemed to be overwhelmed by ones independent of their AHREs, which would undercut a real assessment of stroke/embolic events: there were a total of 49 ischemic strokes (more so in those on placebo: 27 vs 22 on edoxaban), and 42 systemic embolism (also more so on placebo: 28 vs 14 on edoxaban). But these were very small numbers, dwarfed by the cardiovascular outcomes. And this study was terminated early at 21 months, which might not have allowed sufficient time to see if strokes/systemic emboli were a more prominent endpoint

        -- another big concern to me is the efficacy of edoxaban as a once-a-day med. We went through this before with the once-a-day DOAC rivaroxaban (presumably marketed as once-a-day to improve sales over, for example, the twice-a-day apixaban). But a large observational real-world study of Medicare recipients found that rivaroxaban was associated with both more ischemic and  more hemorrhagic events when compared to apixaban, and this was confirmed in a large commercial health care database: https://gmodestmedblogs.blogspot.com/2022/01/atrial-fib-apixaban-outperforms.html

                -- the big issue here is that (somehow) the FDA approved rivaroxaban as a once-a-day med despite pretty shocking drug company shenanigans in their pivotal ROCKET-AF trial (using knowingly inaccurate INR machines for those on warfarin in the comparator group, see above blog for more details on this drug company malfeasance), and, perhaps most significantly, the half-lives of meds are as follows:

                    -- apixaban: elimination half-life of 12 hours (8-15 hours)

                    -- rivaroxaban: 5-9 hours in healthy young subjects  and 11-13 hours in elderly subjects https://pubmed.ncbi.nlm.nih.gov/23999929/)

                    -- edoxaban: 10-14 hours  (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4875962/ )

                        -- so, not much different (ie, should edoxaban be prescribed twice a day since it also has the same elimination half-life as apixaban??). And did the combination of an inappropriate composite endpoint, a shortened study and using an inferior DOAC lead to this negative outcome????

 

Limitations:

-- this study was based on a large number of individuals, two thirds male and 94% white European, which may limit the generalizability of the results more globally

-- the determination of subclinical atrial fibrillation was by stored results of pacemaker or defibrillators, not including loop recorders as being currently used more often

    -- and, there are no great data suggesting that clinical outcomes differ between, for example, the more sensitive loop recorders versus longer-term ECG monitoring

-- they did not assess the true SCAF burden: for the group who had SCAF in the six minute to one hour range, was there a difference if they had that once or twice over a three-year period versus daily?

-- does the currently used CHA₂DS₂-VASc score need to be reworked for patients with SCAF or high AHREs?

-- this group studied were quite sick with high CHA₂DS₂-VASc scores. Would the results apply to those with lower scores but still qualifying for anticoagulation?

-- also, the comorbidities assessed (leading to the high CHA₂DS₂-VASc) were binary. The patient either had them or not. Should the results of CHA₂DS₂-VASc be applied similarly to patients with well-controlled hypertension or diabetes vs ongoing SBP to 180 mmHg and A1c of 12%???  The CHA₂DS₂-VASc is very useful for populations of people, but perhaps should be interpreted differently for those with controlled vs uncontrolled conditions???

 

so, 

-- this study found impressively that patients with at least six minutes of subclinical AF who were on apixaban had significantly fewer strokes/systemic emboli than those on aspirin

-- aspirin, in the study, did not look so great: not only did it not work so well in preventing strokes/systemic emboli than apixaban, but the strokes that happened on aspirin were more disabling than those in the group on apixaban; and the risk of major very severe bleeds was higher on aspirin

-- but, as always, studies beget more studies:

    -- since atrial fibrillation is the most common arrhythmia; and it increases so dramatically by age (from 14 .2 per 1000 person years at age 65-69 to 50.5 per 1000 person-years in those aged 85 and older; and this is only an average wherein some people are at much higher risk than others: https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2801355); and stroke is such an awful disabling or fatal outcome for many people; and since the benefits versus risks of apixaban are so clear when it is indicated:

        -- should we be doing long-term event monitors or even loop recorders on all people who have a high risk profile for stroke (including the noted medical associations as well as some noted sociodemographic ones)? What would be the parameters of determining the high-risk cutpoint?

        -- perhaps patients who have subclinical AF or atrial high-rate episodes identified should then have their CHA₂DS₂-VASc and HAS-BLED scores assessed to further stratify the risks versus benefits of anticoagulation?

            -- One of the studies noted above would suggest that this would be beneficial: https://www.ahajournals.org/doi/epub/10.1161/CIRCULATIONAHA.119.044407

    -- and perhaps most importantly, this article reinforces the importance of aggressively helping all patients improve their lifestyle issues, especially diet/exercise/weight management/stress management/etc, which are associated with so many of the medical risk factors for atrial fibrillation

        -- for example, a study of the Mediterranean diet that emphasized the use of extra-virgin olive oil was associated with a  dramatic decrease in atrial fibrillation incidence: https://gmodestmedblogs.blogspot.com/2014/05/olive-oil-and-atrial-fibrillation.html )

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:

 

 go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

 

if you would like to see the articles in this blog, please email me. 

 

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list