2026 ECNP Congress

PS02-2101

Evaluating the brain age gap in lithium responder and non-responder patients with bipolar disorder using CentileBrain

Elif Özge Aktaş Keskin — E. Şen-Aydın, K. Keskin, M.Ç. Eker · Ege University- SoCAT Lab, Psychiatry, İzmir, Türkiye

2026 ECNP Congress

Background

Bipolar disorder (BD) is linked to a shorter life expectancy and higher rates of age‑related diseases. Research defines BD as a condition of accelerated biological ageing. The “brain-age gap”, derived from structural magnetic resonance imaging (MRI), is a robust biomarker for this process. Lithium is a first-line medication for BD with well-documented mood-stabilising and neuroprotective benefits, such as increased grey matter volume[1,2]. However, it remains unclear whether these properties translate into a measurable reduction in the brain-age gap. This study aimed to compare the brain-age gap between bipolar disorder patients who are lithium responders and non-responders. We hypothesised that responders would exhibit a significantly lower brain-age gap, reflecting an attenuation of accelerated ageing. We additionally explored the influence of the total number of mood episodes and smoking status on this biomarker.

Methods

Forty-four clinically remitted patients with bipolar disorder (HAM-D < 8, YMRS ≤ 4 for ≥1 month) participated. Patients with psychiatric comorbidities or unstable medical conditions were excluded. Using the Alda Scale[3], participants were categorised into responders (n=15; score ≥ 7, ≥1 year therapeutic lithium) and non-responders (n=29; score ≤ 4, no recent lithium). 3T structural T1-weighted MRIs were acquired. Brain age was estimated using the CentileBrain BrainAGE toolbox, benchmarking scans against normative datasets[4]. The dependent variable was the Adjusted Brain-Age Gap (Adjusted_BrainAGE). An Ordinary Least Squares (OLS) regression model assessed lithium response as the primary predictor, with smoking status and total episodes as covariates. Age and sex were not included as additional covariates in the regression model, as they were already accounted for during the estimation of the Adjusted_BrainAGE.

Results

The study sample comprised 44 participants with a mean age of 36.39 ± 8.01 years (range: 23–52 years; median: 35.0 years), reflecting a predominantly young-to-middle-aged adult distribution. Serum lithium concentration was in the therapeutic range for the responder group (0.70 ± 0.15 mmol/L). An ordinary least squares (OLS) regression model accounted for approximately 13.5% of the total variance in the Adjusted_BrainAGE (R² = 0.135, Adjusted R² = 0.046; F(4, 39) = 1.518, p = 0.216). When examining individual predictors, lithium response status was associated with a descriptive reduction in brain age (B = -1.358, SE = 1.064), although this trend did not reach statistical significance (t = -1.277, p = 0.209). Furthermore, the total number of episodes (p = 0.666) and smoking status (p = 0.153) did not significantly contribute to the variance in the brain-age gap.

Conclusions

We found no statistically significant association between lithium response and a reduced brain-age gap, nor did cumulative illness burden (total number of episodes) significantly drive structural brain ageing in this cohort. Although responders showed a descriptive trend towards a lower brain-age gap, statistical power was likely constrained by the limited sample size (N = 44) and predominantly young-to-middle-aged demographic. These findings suggest overall chronicity does not exert a dominant effect on macroscopic structural deviations here. Larger, multicentre cohorts are essential to definitively establish whether the clinical efficacy of lithium translates into a sustained preservation of biological brain age across the lifespan.

References

  1. Vigod, S.N., et al., 2025. Canadian Network for Mood and Anxiety Treatments 2024 Clinical Practice Guideline for the Management of Perinatal Mood, Anxiety, and Related Disorders: Guide de pratique 2024 du Canadian Network for Mood and Anxiety Treatments. Can. J. Psychiatry 70 (6), 429-489.
  2. Lyoo, I.K., Dager, S.R., Kim, J.E., Yoon, S.J., Friedman, S.D., Dunner, D.L., Renshaw, P.F., 2010. Lithium-induced gray matter volume increase as a neural correlate of treatment response in bipolar disorder: a longitudinal brain imaging study. Neuropsychopharmacology 35 (8), 1743-1750.
  3. Grof P, Duffy A, Cavazzoni P, et al. Is response to prophylactic lithium a familial trait? J Clin Psychiatry. 2002;63:942‐947.
  4. Ge, R., et al., 2024. Normative modelling of brain morphometry across the lifespan with CentileBrain: algorithm benchmarking and model optimisation. Lancet Digit. Health 6 (3), e211-e221.