Mental health conditions and lifespan

Estimates of years of life lost come mainly from national registers and specialist health records that follow people after a diagnosis. In a review of reviews, substance use disorders and anorexia nervosa had the highest all-cause mortality ratios of any mental disorder, and the reported reductions in life expectancy were 9 to 24 years for substance use disorders, 13 to 22 years for personality disorders, 10 to 20 years for schizophrenia, 9 to 20 years for bipolar disorder and 7 to 11 years for a depressive episode or recurrent depressive disorder [2]. A Danish register study of 7,369,926 people found mortality rate ratios ranging from 1.92 for mood disorders to 3.91 for substance use disorders, and excess life-years lost ranging from 5.42 years for organic disorders in women to 14.84 years for substance use disorders in men [7].

For schizophrenia, a meta-analysis of 11 studies on five continents estimated an average of 14.5 years of potential life lost (95% confidence interval, the range within which the true value probably lies, 11.2 to 17.8), 15.9 years in men and 13.6 in women. Average life expectancy was 64.7 years, 59.9 in men and 67.6 in women [4]. In Finland, the gap in life expectancy between people with schizophrenia and the general population was 25 years in 1996 and 22.5 years in 2006 [8].

Anorexia nervosa had a weighted mortality rate of 5.1 deaths per 1,000 person-years and a standardised mortality ratio of 5.86 in a meta-analysis of 36 studies, compared with 1.93 for bulimia nervosa; one in five people with anorexia nervosa who died had died by suicide [6]. For ADHD, a matched study of 30,039 adults with a diagnosis in UK primary care estimated a reduction in life expectancy of 6.78 years in men and 8.64 years in women; only about one in nine adults with ADHD in the study population had a diagnosis, so the results may not apply to undiagnosed adults [9].

Anxiety disorders carry a smaller excess risk. In a Danish national register study with more than 30 million person-years of follow-up, people with anxiety disorders had 1.39 times the rate of death from natural causes and 2.46 times the rate from unnatural causes, and co-occurring depression accounted for an important part of the increase [10]. The link between everyday stress, anxiety and health is described under stress and anxiety.

Full deep dive: Mental health conditions ranked by years of life lost.

The largest share of excess deaths is from the same diseases that cause most deaths in the general population. In the 2015 meta-analysis, natural causes accounted for 67.3% of deaths among people with mental disorders [1]. The authors estimated that 14.3% of deaths worldwide, about 8 million a year, are attributable to mental disorders [1]. These deaths are largely missing from standard burden estimates: the Global Burden of Disease Study 2019 counted deaths from mental disorders only for anorexia nervosa and bulimia nervosa, the only mental disorders coded as an underlying cause of death, and its authors noted that its figures do not reflect the premature mortality of people with mental disorders [11].

Suicide accounts for a smaller share of excess deaths but is elevated in several disorders. Borderline personality disorder, anorexia nervosa, depression and bipolar disorder had the highest suicide risks in the review of reviews [2]. In randomised trials of mood disorders, lithium was associated with fewer suicides than placebo (odds ratio 0.13, 95% confidence interval 0.03 to 0.66, across 48 trials and 6,674 participants) [12].

Physical diseases linked to mental disorders, including cardiovascular disease, are influenced by smoking, alcohol use, physical inactivity and access to medical care, all of which differ between people with and without mental disorders. The contribution of smoking and alcohol is described on their own pages.

Depression is the most common mental disorder associated with shortened life. A meta-analysis of 293 studies with 1,813,733 participants from 35 countries found an unadjusted relative risk (the risk in one group divided by the risk in another) of death of 1.64 for people with depression, reduced to 1.52 after adjustment for publication bias. The association was similar in community samples and in patients with heart disease, cancer or kidney disease, and was weaker in longer and higher-quality studies [5]. The review of reviews estimated a reduction in life expectancy of 7 to 11 years for a depressive episode or recurrent depressive disorder [2].

Symptoms below the threshold for a diagnosis are also associated with higher mortality. In 22 studies that measured both, mortality with major depression was not significantly higher than with subthreshold depression (relative risk 1.13, 0.98 to 1.30), and each accounted for an estimated 7% of deaths in the population [13].

Depression is associated with several physical diseases. In 21 studies of initially healthy people, depression was associated with 1.81 times the risk of later coronary heart disease, although the authors concluded that it had not been established as an independent cause because adjustment for other risk factors was incomplete [14]. After a heart attack, depression was associated with 2.38 times the odds of death from any cause and 2.59 times the odds of cardiac death within two years, in 6,367 patients [15]. Depression was associated with a 60% higher risk of developing type 2 diabetes, and diabetes with a 15% higher risk of later depression [16]. Across 28 studies of 317,540 participants, depression was associated with a 45% higher risk of stroke [17].

Suicide risk is raised in depression [2]. Crisis lines provide immediate support to people in suicidal crisis; in the United States the number is 988, by call or text.

Full deep dive: Depression and lifespan.

Depression is associated with changes in several body systems that also drive heart and metabolic disease. People with depression more often show chronic low-grade inflammation and overactivity of the hypothalamic-pituitary-adrenal (HPA) axis, the hormone system that releases cortisol in response to stress, with less consistent evidence for dysregulation of the autonomic nervous system. Depression is also associated with the obesity-related parts of the metabolic syndrome, such as abdominal fat and unhealthy blood lipids. Inflammatory and metabolic changes appear more typical of atypical depression, and high cortisol of melancholic depression [18]. The role of long-term inflammation in ageing is described under inflammation.

Behaviour forms a second pathway. In a meta-analysis of 12 studies, patients with depression had 3.03 times the odds of not following medical treatment recommendations compared with patients without depression, while anxiety showed no consistent association [19]. Smoking is more common among people with mental disorders, and stopping is associated with improved mental health: across 26 studies, quitting was associated with reductions in anxiety, depression and stress compared with continuing to smoke, with effect sizes equal to or larger than those of antidepressant treatment [20].

The main trial evidence comes from the PROSPECT trial, in which 20 primary care practices in three US cities were randomly assigned either to usual care or to two years of depression care management, with a care manager supporting psychotherapy, medication adjustment and monitoring. Among 1,226 older adults followed for a median of 98 months, those with major depression in usual care had 1.90 times the mortality of people without depression, while those with major depression in the intervention practices had 1.09 times the mortality, a difference that was not significant. Compared with usual care, major depression in the intervention practices was associated with 24% lower mortality (hazard ratio, the risk in one group relative to another, 0.76, 95% confidence interval 0.57 to 1.00). No effect was found for minor depression [3].

In severe mental illness, a Finnish register study of 66,881 people with schizophrenia found that 7 to 11 years of cumulative antipsychotic treatment was associated with lower mortality than no antipsychotic use (hazard ratio 0.81, 0.77 to 0.84) [8]. The finding is observational, and people who take treatment consistently may differ in other ways from those who do not.

Several treatments reduce depressive symptoms, although trials of symptoms do not measure lifespan. A network meta-analysis of 522 trials with 116,477 participants found that all 21 antidepressants studied were more effective than placebo, with odds ratios for response from 1.37 to 2.13 [21]. A network meta-analysis of 218 trials with 14,170 participants found moderate reductions in depression with walking or jogging (Hedges' g, a standardised measure of effect size, -0.62), yoga (-0.55) and strength training (-0.49), with larger effects at higher intensity; confidence in the evidence was low for walking or jogging and very low for other forms of exercise [22]. Mind-body forms of exercise are described under yoga and tai chi.

Nearly all estimates of years of life lost are observational. People with mental disorders differ from the general population in income, smoking, alcohol use, housing and access to health care, and these differences can be causes or consequences of the disorder. The review of reviews rated the quality of most included reviews as low [2], and the pooled estimate of excess mortality in depression fell after correction for publication bias and was smaller in higher-quality studies [5].

Reverse causation is also possible: physical illness can lead to depression, so part of the association between depression and heart disease may run from the disease to the mood disorder. Studies of depression and coronary heart disease adjusted incompletely for conventional risk factors and for the severity of existing heart disease [14]. Estimates based on diagnosed cases, such as the ADHD study, may not apply to people who have never been diagnosed [9]. The PROSPECT trial is the main randomised evidence on treatment and mortality, and its result for major depression was of borderline significance [3].