Depression often increases in older adults as they approach the end of life, a phenomenon known as terminal decline. A new study shows that this deterioration in mood accelerates approximately four years before death, and worsens more rapidly in men than in women. This study psychological sciencethis helps explain why historical differences in mental health between the sexes tend to disappear in late adulthood. Tracking these mood changes later in life could ultimately support clinicians in diagnosing and treating mental illness in older adults.
Mental health naturally changes over a person’s lifetime. Average symptoms of depression tend to decrease from middle age to early adulthood. Then, around age 70, this decreasing trend routinely reverses and symptoms begin to increase again. Researchers propose that common age-related changes, such as functional impairment and changes in social roles, are responsible for this late-life increase.
An alternative hypothesis suggests that this late-life increase reflects end-of-life decline. This concept describes the rapid deterioration of various physical and psychological characteristics that occurs shortly before a person’s death. Previous studies have observed declines at the end of life across areas such as memory, overall well-being, and overall life satisfaction.
Andrew Pecas, a neurology researcher at the University of Southern California, and his colleagues wanted to test whether the terminal decline hypothesis applied specifically to depression. They also sought to understand whether this process affects men and women differently. Women typically report depression more often than men throughout most of their lives. In the oldest age groups of the population, these differences often narrow, leaving men and women with similar levels of depressive symptoms. The researchers suspected that a more severe end-of-life decline process in men might explain this narrowing difference. Men also tend to experience a sharp decline in physical health when they are very old, which can make them more vulnerable to mood changes.
To answer these questions, the team analyzed data from the Gene-Environment Interaction Consortium across multiple studies. This is an international research collaboration involving a set of twins from Sweden, Denmark and Australia. The sample included 2,411 community-dwelling older adults who completed at least three longitudinal assessments of their mental health. In this group, 1,491 participants died during the study period, allowing researchers to retrospectively examine their mood before death. Participants completed multiple testing sessions separated by years and kept detailed records of their mental health over time.
Participants completed a questionnaire asking how often they experienced certain emotional or physical symptoms in a given week. The Swedish, Danish and Australian researchers relied on slightly different psychological questionnaires, so the team used statistical methods to harmonize the scores into a single mathematical metric. This standardization process allowed for smooth comparison of results from different countries. Essentially, the researchers cross-adjusted the tests so that a particular score in Denmark meant exactly the same thing as a particular score in Australia.
The researchers then built a statistical tool called the joint model to map the average change in depression symptoms over chronological age. Joint models combine two different types of analysis. One part tracks how certain traits grow or shrink over a timeline. The second part calculates the probability of an endpoint event, such as survival or mortality. This approach allowed the research team to see whether an individual’s mental health trajectory could confidently predict mortality risk.
The results were consistent with expectations for human aging. On average, people showed a small increase in depressive symptoms by age 70, with even larger increases each year after age 70. Statistical models showed a relationship between these late-life changes and mortality. Those who reported a greater annual increase in depression after age 70 had a higher risk of death than those with a more stable mental state. Median survival for those with a significant increase in depression late in life was just over a year and a half shorter than for those with more stable moods.
To verify that impending mortality was indeed driving these results, the team removed all psychological data collected within three years of a participant’s death and ran the model again. Under these circumstances, the rapid rate of decline in mental health after age 70 almost completely leveled off. The association between late-life depression and increased risk of death virtually disappeared, leaving the result not statistically significant. This confirms that the last years of life are the main cause of the spike in depression after age 70.
To pinpoint when this accelerated change begins, the researchers built another type of mathematical model. This is an analytical search for a specific point in time when the rate of change in mood suddenly worsens, working backwards from the time of death. They found that symptoms of depression rapidly accelerate about four years before a person dies. The model documented considerable variation between people, with an average onset of 4 years, but found that some people began terminal decline earlier or later than others.
Comparing the group that died with the group that survived the study period highlighted the reality of this terminal change. Among surviving participants, the rate of change in depressive symptoms remained relatively stable until the final assessment. Over the past four years, the proportion of people who died with symptoms of depression has dramatically increased.
This study also relied heavily on a cotwin control design. Observational studies typically struggle to account for all the invisible variables that can affect a person’s health. Researchers may be able to account for hidden confounders by studying sets of twins in which one child died early and the other survived for at least another four years. Identical twins share all of their genes, and fraternal twins share about half of their genes. Both types of twins typically share a shared childhood environment, including socio-economic upbringing and community support systems.
The researchers analyzed 98 pairs of twins and found that the twins who died had significantly accelerated symptoms of depression compared to their surviving siblings. Because these genetic and childhood factors were equally shared between the twins, the results suggest that terminal decline is caused by factors specifically associated with impending mortality, rather than simply a family history of depression or a harsh childhood.
As hypothesized, the end-of-life decline process looked different for men and women. Men experienced a more severe increase in depressive symptoms after four years compared to women, but women tended to enter this accelerated phase about a year earlier. This large difference in speed means that by the time death occurs, older men have effectively caught up with older women in overall average depression levels. Researchers suggest that the high levels of functional impairment and loss of physical independence that are common in the last years of life may have a stronger negative impact on men than on women.
The research team noted some limitations regarding the population included in the data. Participants from Sweden, Denmark, and Australia were primarily of European descent, limiting the ability to generalize the results to more diverse populations around the world. A broader sample of subjects from different racial and ethnic backgrounds is needed to see if these patterns hold across the world.
The current data did not track specific antidepressant use among participants. The use of therapeutic drugs can alter an individual’s mental health trajectory in ways that cannot be explained by statistical models. Future studies should track drug use to see how medical interventions blunt the severity of terminal deterioration in psychiatric patients. The researchers also were unable to perfectly match the deceased and survivor groups on age and gender, and additional statistical techniques were needed to ensure the main results were robust.
Although this study identifies a timeline of end-of-life decline, it does not reveal the structural causes behind mood changes. Late-life depression may be conceptualized as a depletion of emotional reserve. Emotional reserve represents a type of mental stamina that acts as an emotional buffer that helps people cope with stress. Toward the end of life, people may simply run out of emotional resources, leaving them uniquely vulnerable to psychological distress.
Issues such as decreased physical mobility, loss of friends, diminished sensory awareness, and a growing sense of helplessness can all contribute to mood changes in later life. Further scientific research will be needed to elucidate which specific physical and social losses cause the ultimate increase in psychological distress as humans approach the end of life.
The study, “Terminal Increase in Depressive Symptoms in a Multinational Twin Consortium,” was authored by Andrew J. Petkus, Chandra A. Reynolds, Vibeke S. Catts, Kaare Christensen, Deborah Finkel, Marianne Nygaard, Perminder S. Sachdev, Nancy L. Pedersen, and Margaret Gatz.
