Serum D-Serine Enantiomeric Imbalance and Clinical Progression in Neurodegenerative Disease: A Multi-Diagnosis Cohort Analysis

Background: Predicting which patients with neurodegenerative disease will decline quickly still rests largely on clinical judgement. Serum D-serine, a co-agonist at the N-methyl-D-aspartate (NMDA) receptor whose balance with L-serine reflects glutamatergic and glial activity, is an attractive blood-based candidate, but its prognostic behaviour across neurodegenerative syndromes has not been defined. We prospectively followed 50 consecutive patients with Alzheimer's disease, Parkinson's disease, mild cognitive impairment, amyotrophic lateral sclerosis, Huntington's disease, frontotemporal dementia, or other neurodegenerative disorders for a mean of 18.7 months (range 12–24). Serum D- and L-serine were measured by chiral chromatography (HPLC-DAD, n = 29; LC-MS/MS, n = 21). Associations were tested with Spearman correlation under Benjamini–Hochberg false-discovery-rate (FDR) control, logistic and linear regression, receiver-operating-characteristic (ROC) analysis, and Cox modelling. At baseline, serum D-serine was 2.82 ± 0.70 µmol/L and the D-serine percentage of total serine 2.35 ± 0.47%. The D-serine percentage correlated with activities-of-daily-living impairment (ρ = −0.70), Clinical Dementia Rating (ρ = +0.70) and TNF-α (ρ = +0.66), all surviving FDR control. Serum D-serine rose in 39 of 50 patients during follow-up (median +0.31 µmol/L; P < 0.001) while L-serine was unchanged. Twenty-six patients (52%) progressed. The baseline D-serine percentage did not significantly discriminate progressors from stable patients (AUC 0.612, 95% CI 0.448–0.770; Cox hazard ratio 1.75 per percentage point, 95% CI 0.77–3.96; P = 0.179), whereas neurofilament light chain (NfL) discriminated modestly (AUC 0.724, 95% CI 0.575–0.857). An association that appeared only after covariate adjustment (OR 7.34, 95% CI 1.21–44.55) is unstable with 26 events, and the change in D-serine is coupled to the outcome definition and cannot be interpreted as a predictor. Serum D-serine enantiomeric metrics therefore mirror current disease severity and neuroinflammation; they remain a research-stage candidate rather than a validated prognostic marker.