Family history and genetics

Heritability is the share of the differences between people in a population that is explained by differences in their genes. A study of all 2,872 same-sex Danish twin pairs born between 1870 and 1900 in which both twins survived to age 15, followed until almost all had died, estimated the heritability of lifespan at 0.26 in men and 0.23 in women. The best-fitting model included non-additive genetic effects and environmental factors specific to each individual, and found no evidence of an effect from the environment shared within families [1].

A 2018 analysis of public family trees from Ancestry, covering hundreds of millions of people born in the 1800s and early 1900s, found that correlations in lifespan between blood relatives agreed with the earlier estimates of about 15 to 30%. Most of that correlation, however, was also seen between in-laws, who share no genes. The authors attributed this to assortative mating (people choosing partners who resemble them, here in factors that influence lifespan, whether genetic or environmental), and after accounting for it estimated the true heritability of lifespan at well below 10% [2].

Genetic influence appears stronger at the extreme end of the age range. A study of 20,360 families comprising 314,819 people in the United States and the Netherlands found that survival into the top 10% of a birth cohort (the longest-lived tenth of people born in the same period) was passed on in families, and that people with first- and second-degree relatives in that top 10% had a survival advantage of up to 31%, even when their parents were not long-lived [8]. Other biological processes of aging, such as the shortening of telomeres, are described elsewhere in longevity theories.

Full deep dive: family history.

In the Framingham Offspring Study, 2,302 adults (mean age 44) without cardiovascular disease, whose parents had also been examined in the original Framingham study, were followed for 8 years. Premature cardiovascular disease in at least one parent, defined as starting before 55 in a father or 65 in a mother, was associated with age-adjusted odds of an event 2.6 times as high in men and 2.3 times in women. After adjustment for the offspring's own risk factors, the odds ratios were 2.0 (95% CI 1.2 to 3.1) in men and 1.7 (0.9 to 3.1) in women. Parental disease that began later in life was a weaker predictor [3].

In the same study population, 3,443 stroke-free adults whose parents' strokes had been verified by examination were followed for up to 8 years after each visit. A parental stroke by age 65 was associated with about three times the risk of stroke in the offspring (hazard ratio 2.79, 95% CI 1.68 to 4.66), after adjustment for standard stroke risk factors [9].

Some familial heart disease is caused by single genes. Familial hypercholesterolaemia, an inherited condition causing very high LDL cholesterol from birth, affects an estimated 1 in 500 to 1 in 200 people, or 14 to 34 million worldwide, and carries up to 13 times the risk of coronary heart disease; fewer than 1% of cases are diagnosed in most countries [10]. The European Atherosclerosis Society recommends screening families when a relative has the condition, premature coronary heart disease or sudden premature cardiac death [10]. Heart disease as a whole is described under cardiovascular disease.

The EPIC-InterAct study of 13,869 people in several European countries, including 6,168 who developed type 2 diabetes, found that a family history of the disease was associated with 2.72 times the risk (95% CI 2.48 to 2.99). Adjustment for body mass index, waist size and other established risk factors reduced this only slightly, to 2.44, and a genetic score built from 35 known diabetes-related gene variants explained only 2% of the risk linked to family history. The risk was highest when both parents had diabetes (hazard ratio 5.14) and when a parent was diagnosed before age 50 (4.69) [4].

In the Framingham Offspring Study, 2,527 adults from 1,303 families were followed for 20 years. Compared with those without parental diabetes, type 2 diabetes was 3.4 times as likely with maternal diabetes, 3.5 times with paternal diabetes and 6.1 times when both parents were affected [11].

Family risk of diabetes can be reduced. The Diabetes Prevention Program randomly assigned 3,234 adults with raised blood glucose to placebo, the drug metformin, or a lifestyle programme aiming for at least 7% weight loss and 150 minutes of activity a week. Over an average 2.8 years the lifestyle programme reduced new cases of diabetes by 58% and metformin by 31% compared with placebo [12].

Cancer is caused by changes in genes that control cell growth. Most of these changes arise during life from random copying errors and exposures such as tobacco smoke or ultraviolet light, and up to 10% of all cancers may be caused by inherited genetic changes, such as variants of the BRCA1 and BRCA2 genes that raise the risk of breast and several other cancers [7].

A pooled analysis of 52 studies, including 58,209 women with breast cancer and 101,986 without, found that risk rose with the number of affected first-degree relatives: 1.80 times with one, 2.93 times with two and 3.90 times with three or more. Risk ratios were highest at young ages and when the relative had been diagnosed young. Eight of nine women who develop breast cancer have no affected mother, sister or daughter, and most women with an affected relative never develop the disease [5].

A meta-analysis of 27 studies found that a first-degree relative with colorectal cancer was associated with 2.25 times the risk, more than one affected relative with 4.25 times, and a relative diagnosed before age 45 with 3.87 times [13]. The American Cancer Society recommends that people at average risk start colorectal screening at 45, and that people with a strong family history may need to start earlier, be screened more often or have specific tests [14].

The APOE gene, which makes a protein that carries cholesterol and other fats in the blood and brain, has a major effect on the risk of Alzheimer's disease through its e4 variant. In a study of 7,351 people with Alzheimer's disease and 10,132 without, the lifetime risk of the disease by age 85 was about 11% in men and 14% in women overall, 23 to 30% in people with one copy of e4, and 51 to 60% in people with two copies. The authors noted that these risks resemble those of major single genes such as BRCA1 more than those of the many low-risk variants found in genome-wide studies [15].

Dementia is also linked to modifiable factors. The FINGER trial randomly assigned 1,260 Finnish adults aged 60 to 77 at raised dementia risk to a two-year programme of diet, exercise, brain training and vascular risk monitoring, or to general health advice; cognitive test scores improved slightly more with the programme [16]. The wider set of risk factors is described under dementia risk factors.

Genetic risk for common diseases usually comes from many variants, each with a small effect, which can be combined into a polygenic score. In four studies of 55,685 people, those in the top fifth of a polygenic score for coronary artery disease had 91% higher risk of coronary events than those in the bottom fifth (hazard ratio 1.91). Within every genetic risk group, a favourable lifestyle (at least three of: no current smoking, no obesity, regular physical activity and a healthy diet) was associated with lower risk than an unfavourable one (none or one). Among people at high genetic risk, a favourable lifestyle was associated with 46% lower risk (hazard ratio 0.54, 95% CI 0.47 to 0.63), and in the ARIC study the 10-year rate of coronary events fell from 10.7% to 5.1% [6].

Most of the risk of a first heart attack is linked to modifiable factors. The INTERHEART case-control study of 15,152 people with a first heart attack and 14,820 without, in 52 countries, found that nine factors (abnormal blood fats, smoking, high blood pressure, diabetes, abdominal obesity, psychosocial factors, low fruit and vegetable intake, regular alcohol consumption and lack of regular physical activity) accounted for 90% of the population attributable risk (the share of cases linked to the factors combined) in men and 94% in women [17].

Families share diet, smoking habits, activity levels, income and neighbourhoods as well as genes, so a disease that runs in a family is not necessarily inherited genetically. The lifespan correlation between in-laws, who share no genes, is one sign of this [2], and in the InterAct study known genetic variants explained only 2% of the diabetes risk linked to family history [4]. The Danish twin study, by contrast, found no evidence of a shared family environment effect on lifespan [1], so estimates depend on the method and population studied.

Family histories reported by patients can be incomplete or inaccurate. The Framingham parental studies avoided this by verifying parents' diagnoses through direct examination, which earlier studies relying on offspring reports had not done [3, 9]. Several of the studies above involved mainly people of European ancestry [11, 15], and heritability estimates from historical birth cohorts may not apply to people born today [2].