Depression worsens rapidly in the final four years of life

Older adults often experience a rise in depression as they near the end of their lives, a phenomenon known as terminal decline. A new study reveals that this worsening in mood accelerates roughly four years before death, with men experiencing steeper increases than women. The research, published in Psychological Science, helps explain why historical differences in mental health between sexes tend to vanish in late adulthood. Tracking these late-life mood changes could eventually support clinicians who are diagnosing and treating psychiatric conditions in aging populations.

Mental health naturally shifts across a person’s life span. Average depressive symptoms tend to decrease from middle age into early older adulthood. Then, around age 70, this downward trend routinely reverses, and symptoms begin to climb again. Researchers have proposed that general age-related changes, such as functional impairments or shifting social roles, cause this late-life increase.

Another hypothesis suggests this late-life uptick reflects terminal decline. This concept describes a rapid deterioration in various physical and psychological traits that occurs in close proximity to a person’s death. Previous research has observed terminal decline across domains like memory capability, overall well-being, and general life satisfaction.

University of Southern California neurology researcher Andrew Petkus and his colleagues wanted to test if the terminal decline hypothesis applied specifically to depression. They also sought to understand if this process affects men and women differently. Women typically report more depression than men throughout most of their lives. In the oldest segments of the population, these differences often shrink, leaving men and women with similar levels of depressive symptoms. The researchers wondered if a harsher terminal decline process in men might explain this narrowing gap. Men also tend to experience steeper declines in physical health in very old age, which could make them more vulnerable to shifts in mood.

To answer these questions, the team analyzed data from the Interplay of Genes and Environments Across Multiple Studies consortium. This is an international research collaboration involving sets of twins from Sweden, Denmark, and Australia. The sample included 2,411 community-dwelling older adults who had completed at least three longitudinal assessments of their mental health. Within this group, 1,491 participants died during the study period, giving researchers a way to look retroactively at their mood prior to death. Participants completed multiple testing sessions separated by years, leaving a deep record of their mental health over time.

Participants answered questionnaires asking how frequently they experienced specific emotional or physical symptoms over a given week. Because the Swedish, Danish, and Australian researchers relied on slightly different psychological questionnaires, the team used statistical techniques to harmonize the scores onto a single mathematical metric. This standardizing process allowed them to compare results across the different countries smoothly. Essentially, the researchers cross-calibrated the tests so that a specific score in Denmark meant the exact same thing as a specific score in Australia.

The researchers then built statistical tools called joint models to map out the average change in depressive symptoms across chronological age. A joint model combines two different types of analyses. One part tracks how a specific trait grows or shrinks over a timeline. The second part calculates the probability of an endpoint event occurring, such as survival or mortality. This approach let the team see if a person’s individual mental health trajectory confidently predicted their risk of dying.

The results matched expectations for human aging. On average, people showed small increases in depressive symptoms before age 70, followed by much larger annual increases after age 70. The statistical models showed a relationship between these late-life changes and mortality. Individuals who reported larger annual rate increases in depression after age 70 had an increased risk of death compared to those with more stable mental health. Individuals with profound increases in late-life depression had a median survival time that was a little over a year and a half shorter than individuals with more stable moods.

To verify that impending mortality was actually driving these outcomes, the team removed all psychological data collected within three years of a participant’s death and ran the models again. Under these conditions, the steep rate of mental health decline after age 70 flattened out almost entirely. The association between late-life depression and an increased risk of death was practically erased, leaving a result that was not statistically significant. This confirmed that the final few years of life were primarily responsible for the post-70 depression spike.

To pinpoint exactly when this accelerated change begins, the researchers constructed another type of mathematical model. This one worked backward from the time of death, analytically searching for a specific point in time where the rate of change in mood suddenly worsens. They identified a sharp acceleration in depressive symptoms starting approximately four years before a person died. The models recorded substantial variability from person to person, meaning that while the average onset was four years, some people began their terminal decline earlier or later than others.

Comparing the deceased group to those who survived the study period highlighted the reality of this terminal change. Among the surviving participants, the rate of change in depressive symptoms remained relatively flat leading up to their final assessment. For those who died, the rate of depressive symptoms spiked dramatically after this four-year mark.

The study also relied heavily on a co-twin control design. Observational studies usually struggle to account for all the invisible variables that might influence a person’s health. By studying sets of twins where one died early and the other survived at least four more years, the researchers could account for hidden confounding factors. Identical twins share all their genes, and fraternal twins share roughly half. Both types of twins typically share their early childhood environments, including their socioeconomic upbringing and community support systems.

Analyzing 98 twin pairs, the team found that the deceased twin experienced a much steeper acceleration in depressive symptoms compared to their surviving sibling. Because these genetic and childhood factors were shared evenly between the twins, the results suggest the terminal decline is driven by factors specifically related to impending mortality, rather than just a family history of depression or a tough childhood.

As hypothesized, the terminal decline process looked different for men and women. Men experienced more severe increases in depressive symptoms after the four-year mark compared to women, though women tended to enter this phase of acceleration about a year earlier. This profound difference in velocity means that by the time death occurs, older men have effectively caught up to older women in their overall average depression levels. The researchers suggest that the advanced functional impairment and physical loss of independence common in the last years of life might have a stronger negative impact on men than on women.

The team noted several limitations regarding the populations represented in the data. The participant pools from Sweden, Denmark, and Australia are primarily of European descent, which limits the ability to generalize the results to more diverse populations globally. A broader sample of subjects from varied racial and ethnic backgrounds is required to see if these patterns hold up worldwide.

The current data did not track the specific use of antidepressant medications among the participants. The use of therapeutic drugs could alter an individual’s mental health trajectory in ways the statistical models could not account for. Future studies will need to track medication usage to see how medical interventions might blunt the severity of terminal decline in psychiatric patients. The researchers were also unable to perfectly match the deceased and surviving groups by age and gender, requiring extra statistical techniques to ensure their main results were robust.

While the research identifies the timeline of terminal decline, it does not reveal the structural causes behind the changing mood. Late-life depression might be conceptualized as a depletion of affective reserve. Affective reserve describes a type of mental stamina, acting as an emotional buffer that helps people cope with stress. Towards the very end of life, people may simply run out of affective reserve, leaving them uniquely vulnerable to psychological distress.

Issues like declining physical mobility, the loss of friends, failing sensory perception, or an accumulating sense of helplessness might all contribute to this late-life mood shift. Further scientific inquiry will be needed to untangle which specific physical and social losses drive the final increase in emotional distress as human beings near the end of life.

The study, “Terminal Increases 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.