Paperonium

Cardiology · JAMA Cardiol · 2026

Averaging two biobanks’ risk figures, 1.70 and 1.63, the paper gets 1.88: higher than both.

Its results: the two combinedhigher than bothIts table: one biobankIts table: the other biobankBoth biobanks, combinedby usbetween them1.88×1.70×1.631.661.6×1.7×1.8×1.9×
Its results: thetwo combinedhigher than bothIts table:one biobankIts table: theother biobankBoth biobanks,combined by usbetween them1.88×1.70×1.631.661.6×1.7×1.8×1.9×

Times the risk of an inflamed heart sac, versus people without such a clone

The paper says

People whose blood holds a clone (many cells copied from one mutated cell) had 1.65 times the risk of an inflamed heart sac. Clones with a change in one gene, DNMT3A: 1.88 times. It says these clones explain the link (“accounted for the association”).

Its own data show

That 1.88 combines two biobanks (collections of DNA and health records). Its own table gives each biobank’s figure for these clones: 1.70 and 1.63 times the risk. Its combined row counts exactly the people of those two, no others. Combined the way its methods describe, the two give 1.66.

Combining biobanks this way is a weighted average (the more precise one counts more), and an average always lands between its parts: 1.63 to 1.70. So 1.88 cannot come from them. The claim that these clones explain the link rests on 1.88; at 1.66, they are about as risky as clones in general: 1.65, the paper’s main finding, which stands.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Clonal Hematopoiesis and Risk of Incident Pericarditis and Myocarditis in 2 US Biobank Cohorts · JAMA Cardiol (2026) · doi:10.1001/jamacardio.2026.0258