Aspirin

Secondary prevention, meaning someone who has already had a cardiovascular event, is settled: serious vascular events fell from 8.2 to 6.7 percent per year, a large absolute benefit ^[7].

⚠️ PRIMARY PREVENTION REVERSED, AND THE NUMBERS ARE WORTH SEEING TOGETHER. The 2009 pooled analysis already found the vascular benefit came entirely from non-fatal heart attacks with no effect on vascular death, and called the net value uncertain ^[7]. Then three trials reported. ARRIVE found no difference at all, 4.29 against 4.48 percent, with gastrointestinal bleeding more than doubled ^[8]; its own investigators note the event rate came in so low that it could not answer the moderate-risk question. ASCEND, in people with diabetes, found a real benefit, 8.5 against 9.6 percent, and concluded the absolute benefits were largely counterbalanced by the bleeding hazard ^[9]. ASPREE, in nearly 20,000 healthy older adults, found no effect on disability-free survival, no significant cardiovascular benefit, a 38 percent increase in major haemorrhage, and an unexpected excess of deaths ^[10][11].

Pooled across 164,225 participants: a number needed to treat of 265 against a number needed to harm of 210 ^[12]. The guidance followed, and now recommends against starting aspirin for primary prevention at 60 and over ^[13].

⚠️ THE CANCER QUESTION IS UNRESOLVED IN BOTH DIRECTIONS, which is not the same as promising. Pooled trial data reported a 20-year reduction in cancer death, strongest for gastrointestinal cancers, and a benefit in Lynch syndrome at 600 mg daily ^[19]. Against that, ASCEND found no difference in cancer at all, and ASPREE reported a 31 percent increase in cancer death and a signal that cancers presented at a later stage ^[20]. The one thing that replicates everywhere is the bleeding.

The bleeding risk is the one finding that has replicated in every dataset, and it is the reason the primary-prevention case fell apart.

Gastrointestinal bleeding roughly doubles. In primary prevention the adjusted relative risk of upper gastrointestinal bleeding is 1.90 against 1.40 in secondary prevention ^[15]. ⚠️ Counterintuitively the relative risk is HIGHER in primary prevention while the absolute harm is lower, because the baseline risk differs: about one extra bleed per 601 person-years in primary prevention against one per 391 in secondary.

Bleeding into the head is rarer and worse. Across 13 trials in people without known cardiovascular disease, low-dose aspirin raised any intracranial bleeding by 37 percent, about two extra bleeds per 1,000 people, with the largest relative increase in subdural and extradural bleeding ^[14]. The risk was heightened in people of Asian background and at low body weight.

⚠️ The gastric harm is systemic, not a local burn. Inhibition of stomach prostaglandins tracked blood COX-1 potency, not anything topical, which is why no anti-inflammatory of this class spares the stomach at working doses.

Reye's syndrome is the reason aspirin is not given to children. National surveillance recorded 1,207 cases with an overall case fatality of 31 percent, and the collapse from 555 cases in 1980 to under 36 a year after 1987 followed the salicylate warnings directly ^[16]. It is now so rare that a suspected case should be investigated for treatable metabolic disorders that mimic it.

Aspirin-exacerbated respiratory disease affects about 7 percent of adults with asthma and about 15 percent of those with severe asthma ^[17]. It is not an allergy: reactions happen on first exposure, depend on COX-1 inhibition rather than chemical class, and cross-react to every COX-1-inhibiting anti-inflammatory.

Willow bark was used for fever and pain for millennia; salicylic acid was isolated in the nineteenth century and was effective and brutal on the stomach. Acetylating it, done at Bayer in 1897, produced a tolerable version and one of the first true blockbuster medicines. Its mechanism went unexplained for seventy years until the prostaglandin work of the early 1970s, and the acetylation of the enzyme itself was pinned down in 1975 and visualised structurally in 1995. Its second life as a heart drug began when it became clear that platelets, uniquely, could not recover from it. Its third act has been a retreat: the same trials that confirmed its value after a heart attack showed that giving it to healthy people trades a small number of prevented events for a similar number of caused bleeds.