aspirin increases risk of heart failure

 

A recent article documented the increased risk of heart failure in people on aspirin therapy (see aspirin inc chf ESCHF2021 in dropbox in dropbox, or DOI: 10.1002/ehf2.13688)

 

Details:

-- 30,827 patients at risk for heart failure (HF) were in 6 observational studies and were on aspirin at baseline (ASCOT, FLEMENGHO, HEALTH ABC, HULL LIFELAB, PREDICTOR and PROSPER studies). For this study, these patients were divided into a derivation set of data from ASCOT, and validation set from the other 5 studies

-- mean age 69, 34% women

-- 22% diabetic, 26% current smokers, 67% alcohol consumption, BP 157/89 mmHg, HR 70, BMI 28, total cholesterol/HDL ratio 4.6

-- MI in 3%, coronary heart disease 26%, stroke 10%, atrial fibrillation 1%; 25% on aspirin (total of 7698 people on aspirin)

-- follow-up 5.3 years (2.1-11.7 yrs)

--1330 patients developed HF in the course of the study

-- Primary outcome: composite of fatal and non-fatal HF

 

Results:

-- HF incidence: 26% increased, HR 1.26 (1.10-1.46), p<0.001, adjusting mathematically for sex, age, BMI, smoking, alcohol, blood pressure, heart rate, total cholesterol/HDL ratio, creatinine, use of renin-angiotensin/RAAS inhibitors, calcium channel blockers, diuretics, b-blockers, and lipid-lowering treatment

-- using propensity-score-matching (a statistical technique to equalize the groups in order to make a fairer comparison, see below):

    -- HF incidence overall: 26% increased, HR 1.26 (1.10-1.44), p<0.001, in the fully adjusted model above

    -- HF incidence in patients without history of cardiovascular disease: 27% increase, HR 1.27 (1.10-1.46), p=0.001, in the fully adjusted model above

-- assessing participants developing HF within 2 years of enrollment: 23% increase, HR 1.23 (1.06-1.41), p=0.004

-- HF risk was consistent across strata of systolic and diastolic BP, and higher in those not on diuretics as well as those who were on statins; and was also higher risk in men and those >69yo

-- comparing the findings in patients to assess the role of a history of cardiovascular disease, in their validation cohort:

    -- Analysis controlling for all of the factors above (their Model 3): HR 1.39 (1.22-1.57), p<0.001 [ie, would include those without and with history of heart disease, the primary and secondary prevention groups]

    -- Analysis controlling for all of the factors above plus history of cardiovascular disease (their Model 4): HR 1.17 (1.03-1.34), p=0.02 [ie mathematically eliminating the role CVD plays]

 

Commentary:

-- in the past, we more readily prescribed aspirin for both primary cardiovascular disease and colorectal cancer prevention, and this got a "B" rating for those 50-59yo by USPSTF (see https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/aspirin-to-prevent-cardiovascular-disease-and-cancer , for the 2016 recommendations). But more recent trials have suggested minimal or no overall benefit for aspirin when assessing both the long-term cardiovascular protection plus the short-term risks of major bleeding. And the USPSTF has new draft recommendations as of Oct 2021 suggesting this B rating has dropped to a C (see https://www.uspreventiveservicestaskforce.org/uspstf/draft-recommendation/aspirin-use-to-prevent-cardiovascular-disease-preventive-medication ), grade deflation….

 

-- this study found a 26% increased HF incidence overall in those on aspirin, with a still-significant but less impressive increase in those when controlling for underlying coronary vascular disease [underlying heart disease does predispose people to HF]

-- this increase in HF from aspirin really is not so surprising, since NSAIDs are associated with heart failure (see nsaids heart failure bmj2016 in dropbox or doi.org/10.1136/bmj.i4857; or nsaids cardiovasc risk cureus2017 in dropbox or DOI: 10.7759/cureus.1144)

-- there were prior theories that aspirin or anticoagulation might improve heart failure, since HF is considered a prothrombotic state as well as one with chronic inflammation: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6439138/ , but 2 trials (WATCH and WARCEF in 2009) did not find benefit of aspirin (presumably targeting the inflammatory component) over just the antithrombotic benefit of warfarin; the WARCEF trial, for example, assessed people in sinus rhythm but HF with reduced ejection fraction and found aspirin and warfarin were equivalent). Overall, the relationship between aspirin and HF has not been rigorously studied. A Danish study of 12,277 patients with new-onset HF but normal sinus rhythm did find 25% increased HF readmissions in the group on aspirin but no difference in composite of all-cause mortality or MI or stroke admission (see DOI: 10.1016/j.jchf.2017.09.021 ). And the WATCH trial did find that aspirin vs warfarin did find significantly higher hospitalization rates in the aspirin group (218 vs 155 on warfarin, p<0.001)

-- an analysis comparing aspirin’s benefits vs harms in primary prevention in the elderly (which itself is a major heart disease risk factor) did not establish benefit (see http://gmodestmedblogs.blogspot.com/2018/09/aspirin-in-healthy-adults-harm-without.html )

-- secondary prevention studies: there are even studies questioning the use of aspirin in those with documented coronary artery disease, noting the possibility of stopping aspirin earlier in those getting percutaneous coronary interventions on dual antiplatelet therapy, or not using aspirin with anticoagulation in those with atrial fibrillation and chronic coronary syndromes: see https://www.ahajournals.org/doi/epub/10.1161/CIRCULATIONAHA.120.045695 ; there is increasing evidence that those with CAD and atrial fibrillation do better overall on just a DOAC without the added aspirin (see http://gmodestmedblogs.blogspot.com/2019/09/antiplatelet-plus-anticoag-in-cad-with.html)

-- the analysis above limiting HF to the first 2 years does support their conclusions: it makes it more likely that aspirin is the culprit and not just progressive heart disease over time

 

Limitations:

-- there clearly is a gradation of CAD risk in primary prevention studies, as reflected in the risk calculators. Is someone who smokes, has undertreated hypertension/diabetes/hyperlipidemia but no documented cardiovascular event really that different from someone with documented CAD?? the former patient is at hugely high risk.  On the other hand, is someone with stable CAD who has lowered their blood pressure/stopped smoking/aggressively treated their hyperlipidemia benefiting from aspirin enough to justify the short-term risks of a bad bleed?? I would posit that our binary classification (secondary vs primary prevention) may obscure some important individual patient differences in future cardiovascular risk and lead to inappropriate patient stratification

-- are there cohorts of patients with CAD on aspirin who should not be on aspirin because of having difficult-to-treat heart failure? what is the threshold where we should stop aspirin therapy in those who would otherwise qualify; as a perspective on the magnitude of the benefit of aspirin in the secondary prevention studies, aspirin seems to decrease the risk from 8.2% to 6.7% per year (eg, see https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(09)60503-1/fulltext ). This brings up the typical clinical conundrum: medical studies reductionistically (if that is a word....) take their large number of patients studied, then devolve them statistically into the single average person who is xx years old/xx% white/ xx% female/ with xx% of the various medical and psychosocial comorbidities. is this median person in any way like the 89 yo Capeverdean woman in front of me who has an eGFR of 31? Is she going to benefit from treatment xxx as found in the studies, studies that had zero Capeverdean women (and their unique cultural issues), zero patients over 85yo, and had excluded everyone with kidney failure???  Of course, we need to make clinical decisions based on best guesses, and we do that all the time, but best guesses are not always correct….. [this perhaps nihilistic view is moderated by having several studies done by different researchers in different cohorts of patients and preferably in different countries]

-- and, another statistics issue: propensity-score-matching is used to make the dissimilar groups of patients being compared into more similar ones. It does make sense to assess outcomes with the 2 groups being as similar as possible in terms of demographics and pre-existing problems, but this does not always work: eg see http://gmodestmedblogs.blogspot.com/2020/03/tramadol-fo-oa-inc-mortalityprobs-with.html

-- the increased HF risk in those on aspirin plus on statins found in this study likely is related higher levels of more severe heart disease leading them to being on statins, and this underlying heart disease would itself predispose them to heart failure. and the use of diuretics may decrease the presentation of heart failure by effectively treating early heart failure

-- studies like this one, which combine different studies with different inclusion/exclusion criteria and different methodologies/assessments, do not have close to the statistical rigor of a randomized controlled trial. The fact that the researchers in the study established their initial data from a large single study (ASCOT, the derivation set), then found essentially the same results in their 5-study validation set, helps support the overall validity of their conclusions. And the fact that other prostaglandin-targeted NSAIDs have found similar results as in this aspirin analysis further supports their conclusions.  BUT, the limitations of this kind of study can lead to wrong conclusions (eg: as with prior studies on post-menopausal estrogens and heart disease, or vitamin therapy for hyper-homocysteinemia and vascular disease,…)

 

So, pretty clear association in this retrospective analysis between aspirin use and incident heart failure, positioning the potential long-term benefit of aspirin therapy against the short-term risks of heart failure (and increased bleeding).  And, of course, one of the issues for the patient is that the short-term risks may be quite traumatic and definitive (intracerebral bleed) vs the unknown but likely higher degree of long-term protection from the aspirin. But, in the "do no harm" mantra, we really need more studies to be sure (as editorialists say about almost every study...)

 

geoff

 

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

 

For access to the dropbox, go to link: https://www.dropbox.com/sh/0bmvtita8mzms11/XDTwHySFFg

Then go to "clinic", then to "clinical stuff" for articles, or go to https://www.dropbox.com/sh/nyle22q1fn6lkpk/AAB9B2hBj5Kw4gtrJAkI-UF8a?dl=0 for the powerpoint presentations

 

to get access to all of the blogs (2 options):

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

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

3. 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

 

or: go to https://www.bucommunitymedicine.org/ , a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs (though there have been a few that did not upload over the last year or two). but overall it is much easier to view blogs and displays more at a time.

 

 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list