Grade A
What the trials found
Grade A means someone ran the experiment on that outcome. A trial that finds nothing is still A. We are not claiming a protocol works. We are putting every trial we have next to the score so you can decide.
None of the eight scored metrics is scored on Grade A evidence. Blood pressure and HbA1c each have trials on this page; the engine still runs on cohorts (Lewington, Khaw). Nobody has randomized people to walk 7,000 steps a day and counted deaths. That trial is unaffordable and unblindable.
How a letter is assigned is on How we grade evidence. The eight home headlines are all Grade B — large groups followed for years, not trials.
Did not cut deaths — or raised them
A trial that finds nothing is still Grade A. The letter is how hard the claim was tested, not whether the result was flattering.
- Grade APahor et al. 2014, LIFE trial
Walking and exercise cut disability. The authors wrote that they did not decrease the death rate.
Older adults were randomized to a physical-activity program or to education. Major mobility disability fell. Mortality data were inconclusive. Nobody has run the equivalent trial on daily steps.
Gait Speed — not one of the eight scored metrics
- Grade ABrunström & Carlberg 2018 — 74 trials, 306,273 people
Treating blood pressure that was already below 140 did not change all-cause mortality.
In people without prior heart disease, death rates fell when treatment started from 140 and above. Below that baseline, deaths were statistically unchanged. Do not read the SPRINT or Bundy results as “push an already-normal cuff lower and people live longer.”
Blood Pressure — scored on a cohort, not this trial
- Grade AJaeger et al. 2022 — SPRINT extended follow-up
SPRINT’s death benefit faded once intensive control stopped.
During the trial, the intensive arm had fewer deaths. After a median 8.8 years — once that intensity lapsed and outpatient readings drifted back toward 140 — the intensive arm no longer had fewer deaths. A marker, not a one-time treatment effect.
Blood Pressure — scored on a cohort, not this trial
- Grade ARahimi et al. 2021 — 48 trials, 344,716 people
Heart attacks and strokes fell about 10% per 5-point treated drop. All-cause deaths barely moved.
About 2% fewer deaths per 5 points, and the range includes no change. Treatment moves cardiovascular events more reliably than it moves death from any cause.
Blood Pressure — scored on a cohort, not this trial
- Grade AACCORD (Gerstein et al. 2008)
Aiming under 6% A1c in high-risk type 2 diabetes produced more deaths, not fewer.
The intensive arm had about 22% more deaths. Heart attacks and strokes did not fall enough to count. A target this calculator scores as ordinary raised death in that trial. Do not chase it yourself.
HbA1c — scored on a cohort, not this trial
- Grade AVITAL (Manson et al. 2018) — 25,871 US adults
A gram a day of marine omega-3 did not reduce all-cause deaths.
The trial missed its primary heart endpoint. Total heart attacks were lower on a secondary look. Supplementing does not reproduce the mortality pattern seen in blood-level cohort studies.
Omega-3 Index — not one of the eight scored metrics
- Grade AASCEND (Bowman et al. 2018) — 15,480 adults with diabetes
Same fish-oil dose, people with diabetes, no change in all-cause deaths.
Neither the primary vascular endpoint nor all-cause mortality differed from placebo. Two large trials, same answer.
Omega-3 Index — not one of the eight scored metrics
Fewer deaths — in treated, higher-risk people
These are treatment trials, mostly on high blood pressure already being treated. They are not the home-cuff curve the calculator scores.
- Grade ABundy et al. 2017 — 42 trials, 144,220 treated patients
People who landed at 120–124 on treatment had about 27% fewer deaths than those who landed at 130–134.
That is a treated, achieved number in people already being treated for high blood pressure — not a home cuff sitting at 120 in someone who was never high.
Blood Pressure — scored on a cohort, not this trial
- Grade AEttehad et al. 2016 — 123 trials, 613,815 people
Every 10-point drop in treated systolic pressure came with about 13% fewer deaths.
A treatment slope from randomized trials. Not the observational usual-blood-pressure curve the calculator scores, and not a licence to treat an already-normal reading.
Blood Pressure — scored on a cohort, not this trial
- Grade ASPRINT (Wright et al. 2015) — 9,361 high-risk adults without diabetes
Target under 120 vs under 140: about 27% fewer deaths during the trial.
Stopped early at a median 3.26 years. High-risk adults, no diabetes. Jaeger’s later follow-up is the same trial after that intensity lapsed.
Blood Pressure — scored on a cohort, not this trial
- Grade AKhan et al. 2019 (treated drop) / Welsh et al. 2019 (prediction)
Treating ApoB down 10 mg/dL tracked with about 5% fewer deaths.
Adding ApoB on top of standard cholesterol barely improved prediction. ApoB is not one of the eight scored metrics. Marker, not every way of moving the number.
ApoB — not one of the eight scored metrics
A positive result that does not travel
One high-dose prescription drug cut heart attacks and strokes in a high-risk group. That is not a fish-oil capsule, and it is not all-cause death.
- Grade AREDUCE-IT (Bhatt et al. 2018) — 8,179 statin-treated patients
A high-dose purified EPA drug cut heart attacks and strokes about 25%.
Cardiovascular death also fell. This is a prescription-only, high-dose EPA-only drug in people with high triglycerides, most of whom already had heart disease. It is not a typical fish-oil capsule, and it is not an all-cause death result.
Omega-3 Index — not one of the eight scored metrics
Full bibliography and model limitations: Methodology & Sources.