Counts describe this library, not the strength of evidence.
How to read this evidence
Separate prevention outcomes, risk-factor changes and screening performance. A drug trial in established disease, an exercise blood-pressure analysis and a smartwatch alert study cannot be treated as equivalent evidence for a healthy individual.
Across seven observational studies with 109,365 participants, a weaker day-night activity contrast (lower relative amplitude) was associated with higher cardiovascular mortality (hazard ratio [HR] 1.50, 95% CI 1.13 to 1.98) and all-cause mortality (HR 1.46, 95% CI 1.27 to 1.68). The evidence base is small and observational, so it cannot show that improving the rhythm lowers risk, and illness may itself disturb the rhythm.
In a US national survey of 16,268 adults, weekend catch-up sleep (weekend minus weekday sleep) was not significantly associated with hypertension overall (p>0.05). Only the 0-1 h group had lower odds than the group with 0 h or less (OR 0.814, 95% CI 0.666-0.995), and this cross-sectional result needs confirmation.
Among 106 male endurance athletes aged 50 or older, 25 (23.5%) had ventricular arrhythmia on implanted loop recorders over a median 796 days, and myocardial fibrosis was found in 19 (76.0%) of them vs 31 (38.3%) of the rest. The group is small and highly selected, so it says little about other athletes or ordinary exercisers.
In 233 healthy adults aged 60-84 doing six months of home high-intensity interval training (HIIT), greater total, frequency and intensity adherence was linked to larger VO2peak gains (total adherence SF 0.70, interval 0.19-0.93), but duration adherence and lactate threshold changes were not. This is an association within a trial, so people who stick to plans may differ in other ways.
Pooling 15 randomised trials with 334 sedentary adults, moderate-to-low intensity exercise snacks were linked to lower fasting plasma glucose (SMD -0.52, 95% CI -0.93 to -0.12) and LDL cholesterol (SMD -0.51, -0.84 to -0.18), but not to changes in fasting insulin or HDL cholesterol. The evidence base is small, and the subgroup findings are exploratory.
In 12,743 people who had a hospital treadmill stress test, systolic blood pressure divided by peak METs was linked to cardiovascular events; people in the top quarter had hazard ratios of 1.60 to 2.54 across test stages, while raw exercise systolic pressure alone was not linked (P > .05). It is an observational record-linkage study, so it shows risk association, not that lowering the ratio prevents events.
In 429,740 people with type 2 diabetes in Hong Kong, systolic blood pressure above 120-129 mmHg and diastolic above 70-79 mmHg were linked to progressively higher risks of cardiovascular disease, kidney disease, kidney failure and death, with the strongest associations in the youngest adults. This is an observational cohort, so it shows association rather than that lowering pressure prevents these events.
In a randomized trial of 62 people with hypertension, 8 weeks of AI-assisted remote rehabilitation improved 6-minute walk distance by 62.77 more than conventional offline exercise guidance (95% CI 26.33-99.22; P=.002) and lowered systolic blood pressure by 4.11 more (P=.046). The trial was small, open-label and short, and the systolic result barely reached significance.
Among 71,542 French women (845 with Parkinson's disease), high adherence to the Mediterranean or MIND diet was not linked to Parkinson's overall, but was linked to lower incidence under age 71 (HR about 0.75 for both). The confidence intervals reach 1.00 for one diet, and the finding comes from a subgroup, so it is suggestive.
Among 106 asymptomatic male cyclists and triathletes aged 50 or older, 50 (47.2%) had myocardial fibrosis on MRI, and fibrosis was linked to a higher hazard of ventricular arrhythmia on a loop recorder (HR 4.7, 95% CI 1.8-12.8) over a median 720 days. Few men, short follow-up and a wide interval make this a signal, not a risk calculator.
Among 36 251 UK Biobank participants followed for a median of 13.9 years, a rise in visit-to-visit systolic blood pressure variability was associated with a 23%-33% higher risk of cardiovascular disease, coronary heart disease, stroke, chronic kidney disease and death. It is an observational association, and the readings were taken from health records before enrolment, not from a standard protocol.
An AI model reading routine ECGs (AIRE-HTN) predicted new-onset hypertension with a C index of 0.70 (95% CI 0.69-0.71) in both a Boston hospital cohort and the UK Biobank. That is moderate discrimination in two non-general-population cohorts, so it is a research tool, not a screening test.
In 104,867 Copenhagen adults followed for a median 9.2 years, remnant cholesterol explained 15% (95% CI 9.7-20%) of the excess heart attack risk from current smoking, 20% (13-27%) from low physical activity and 12% (6.6-18%) from low adherence to dietary guidelines. As a mediation analysis of observational data, it cannot show what lowering remnant cholesterol would do.
In 648 older adults (mean age 69.88), higher cardiorespiratory fitness (CRF) was linked to better performance in all five cognitive domains tested, after controlling for covariates. This is a snapshot at baseline, so it cannot show that building fitness improves cognition, and the sample was 71.1% women.
Among 1,408 American Indian adults in the Strong Heart Family Study, those with depressive symptoms at baseline (CES-D 16 or higher) had 54% higher odds of developing hypertension later (OR 1.54; 95% CI 1.06-2.23). This is an observational association from one cohort, so it cannot show that depression causes high blood pressure.
In 39 older adults with hypertension on medication, systolic blood pressure fell by 13 mmHg with resistance training and 12 mmHg with high-intensity interval training (HIIT) at 8 weeks (both P<0.05), and fewer people were classed as non-responders. With 13 people per group and responders defined statistically, the individual response labels are uncertain.
In 19 apparently healthy adults, the Garmin fēnix 6 VO2max estimate agreed well with a laboratory metabolic system (30 s averaged data: MAPE 7.05%, concordance correlation 0.73), but its blood oxygen readings in 22 adults did not (concordance correlation 0.10 for combined conditions). The samples were small and only one watch model was tested.
Pooling 148 randomized trials (8673 participants), exercise training lowered LDL cholesterol by 7.22 mg/dL and triglycerides by 8.01 mg/dL and raised HDL by 2.11 mg/dL versus sedentary controls, modest changes of 3.5-11.7%. Prediction intervals suggest improvement appeared in only one-quarter of individual studies, so effects vary widely.