Reading the MADRS Longitudinally: Three Trajectories That Predict Mortality in Older Adults
- SNAC-K cohort study (n=2,118 community-dwelling adults aged ≥60, Sweden): using MADRS at baseline and up to 3 follow-ups over 9 years identified 3 distinct depression trajectory classes — low-stable, increasing, and U-shaped.
- Increasing trajectory risk factors: high financial strain, poor social support, slow gait speed; U-shaped trajectory risk factor: multimorbidity. These predictors are trajectory-specific — not interchangeable.
- The increasing trajectory was independently associated with significantly higher all-cause mortality after adjustment for age, sex, multimorbidity, and functional status.
- Gait speed emerged as a simple, no-cost clinical marker that predicted depression trajectory classification — slower gait at baseline correlated with increasing depressive symptoms over 9 years.
When we assess depression in older adults, we typically capture a cross-sectional snapshot. A score above cutoff means depression; below it means not. But depression in late life is rarely static — and this Swedish cohort study reframes MADRS as a longitudinal instrument, showing that the trajectory matters more than any single score for predicting who will deteriorate and who will not.
The Three Trajectories in Clinical Terms
The SNAC-K study followed 2,118 dementia-free community-dwelling adults aged 60 and older with MADRS assessments repeated over nine years. Growth mixture modeling identified three depression trajectory classes:
Low-stable (majority): consistently low MADRS scores throughout follow-up. This is the expected clinical course. No special concern.
Increasing: MADRS scores start in a low-moderate range and rise steadily over the observation period. This trajectory is predicted by two modifiable and two quasi-modifiable factors: poor social support and high financial strain (modifiable), slow gait speed and low socioeconomic status (harder to change but important to monitor). This trajectory is independently associated with significantly higher all-cause mortality, even after controlling for confounders.
U-shaped: MADRS scores are elevated at baseline, decrease in mid-follow-up, then rise again. This trajectory is predicted by multimorbidity — the accumulation of physical health conditions. Patients cycle through symptom exacerbation and partial remission in parallel with their medical load.
For Your Practice
The practical value here is in triage, not diagnosis. Three questions at a geriatric intake can meaningfully stratify who is at elevated risk for a deteriorating depression course:
- Gait speed — formally (4-meter walk test) or observationally. Slow gait is a proxy for systemic physiological decompensation and predicts increasing depression independently of other factors.
- Social support quality — not just presence of a caregiver, but subjective social connectedness. The SNAC-K study distinguished between social network size and perceived support quality; it's the latter that predicts trajectory.
- Multimorbidity count — patients with three or more concurrent physical conditions are at risk for a U-shaped course; plan for recurrence, not just remission.
Use MADRS serially in older patients rather than as a one-time screen. A score that is modestly elevated but rising over two assessment points in a patient with slow gait and poor social support is a high-risk profile regardless of the absolute number. The patient who scores 12 on first assessment and 20 on second is more concerning than one who scores 24 at both time points.
In geriatric depression, a rising MADRS tells you more than a high MADRS — trajectory predicts mortality where a single cutoff cannot.
Observational design precludes causal inference. SNAC-K is a Swedish urban cohort; cultural and healthcare context differs from other settings. MADRS was not specifically designed for serial self-report in community settings.