Depression as a Signal of MS Progression

Despite advances in imaging and therapeutics, disability accumulation remains difficult to predict in progressive multiple sclerosis (MS). A study published in the Journal of Neurology positions commonly collected patient-reported depression and anxiety symptoms as potential early indicators of disease progression. Drawing from a well-defined cohort of individuals with primary progressive MS (PPMS), the investigators examined whether scores from the Hospital Anxiety and Depression Scale (HADS) could anticipate worsening disability across multiple functional domains.

Study Design

Data was drawn from the Dutch multicenter SPIN-P study, focusing exclusively on adults with PPMS. The study included 111 participants with at least one year of follow-up, of whom 62 were followed for two years. Depression and anxiety symptoms were assessed using scores from HADS questionnaires. Disability progression was evaluated at both one and two years using three predefined approaches:

• Single-domain measure:
• Expanded Disability Status Scale (EDSS)
• Three-domain composite endpoint:
• EDSS
• Timed 25-Foot Walk (T25FW)
• Arm Function in Multiple Sclerosis Questionnaire (AMSQ)
• Five-domain composite endpoint:
• EDSS
• T25FW
• AMSQ
• Symbol Digit Modalities Test (SDMT)
• Patient-Determined Disease Steps (PDDS)

To examine predictive relationships, the investigators applied logistic regression models adjusted for age, alongside Boruta feature selection.

Results

A central observation is that disability progression tends to occur within individual functional domains rather than across multiple areas simultaneously. Overlap between different measures of progression is limited, highlighting the heterogeneous and compartmentalized nature of functional decline in this population.

At one year, higher total HADS scores predicted progression on the five-variable composite endpoint (OR 3.10, p=0.05), although predictive performance was modest (AUC 0.59). Anxiety and depression sub-scores showed directional but non-significant trends.

By two years, the signal sharpened:

• Depression sub-score:
• EDSS progression (OR 8.88, p=0.04)
• 3-variable composite (OR 7.20, p=0.03)
• 5-variable composite (OR 9.46, p=0.02)
• Total HADS:
• EDSS progression (OR 11.63, p=0.03)
• 5-variable composite (OR 5.56, p=0.04)

Notably, anxiety did not emerge as a consistent independent predictor, suggesting that depressive symptomatology may more closely reflect underlying disease processes.

Median baseline scores further support this pattern. At two years, individuals with progression on the five-variable composite had higher depression scores (median 7 vs. 3, p=0.01) and total HADS scores (median 11 vs. 7, p=0.01).

From Questionnaire to Biomarker?

The study goes a step further by dissecting individual HADS items using Boruta feature selection, a machine learning technique that identifies the most informative predictors. Two depression-related items stood out:

• “I feel as if I am slowed down”
• “I feel cheerful”

Together, these items achieved predictive performance comparable to the full questionnaire (AUC 0.70). This raises the possibility that subjective mood states may reflect shared neurobiological pathways with MS progression. Proposed mechanisms include immune-mediated processes and structural brain changes linking depression and neurodegeneration, although causality remains unresolved.

Clinical Interpretation and Limitations

Limitations include sample size, incomplete data for some functional measures, and reliance on a screening tool rather than diagnostic psychiatric evaluation. Nonetheless, the consistency of findings across endpoints and timepoints strengthens confidence in the observed relationships.

Importantly, the study highlights that relying solely on EDSS may underestimate progression. Only about 15% of progressors worsened exclusively on EDSS, reinforcing the value of composite and patient-reported measures. While the associations are consistent, predictive accuracy remains modest. HADS scores alone are unlikely to serve as standalone prognostic tools. Instead, they may function as complementary biomarkers, augmenting clinical, imaging, and laboratory data.

Reference:
Klein Kranenbarg RAM, Blok KM, van Hasselt Y, et al. Symptoms of depression and anxiety are early predictors of multi-domain disability progression in progressive MS. J Neurol. 2026;273(3):182. doi:10.1007/s00415-026-13701-8