Postmortem evidence of decreased brain pH in major depressive disorder: a systematic review and meta-analysis

Abstract

Introduction: Major depressive disorder (MDD) is a prevalent and debilitating mental disorder that shares symptoms, genetics, and molecular changes in the brain with other psychiatric disorders, such as schizophrenia and bipolar disorder. Decreased brain pH, associated with increased lactate levels due to altered energy metabolism and neuronal hyperexcitation, has been consistently observed in schizophrenia and bipolar disorder. We recently demonstrated similar brain alterations in various animal models of neuropsychiatric disorders, including MDD. However, our understanding of brain pH alterations in human patients with MDD remains limited.

Methods: We conducted meta-analyses to assess postmortem brain pH in patients with MDD compared to control subjects, examining its relationships with recurrence of depressive episodes and illness duration, utilizing publicly available demographic data. Studies reporting individual raw pH data were identified through searches in the Stanley Medical Research Institute database, NCBI GEO database, PubMed, and Google Scholar. The data were analyzed using the random effects model, ANOVA, and ANCOVA.

Results: The random effects model, using 39 curated datasets (790 patients and 957 controls), indicated a significant decrease in brain pH in patients with MDD (Hedges' g = -0.23, p = 0.0056). A two-way ANCOVA revealed that the effect of diagnosis on pH remained significant when considering covariates, including postmortem interval, age at death, and sex. Patients with recurrent episodes, but not a single episode, showed significantly lower pH than controls in both females and males (256 patients and 279 controls from seven datasets). Furthermore, a significant negative correlation was observed between brain pH and illness duration (115 patients from five datasets). Female preponderance of decreased pH was also found, possibly due to a longer illness duration and a higher tendency of recurrent episodes in females.

Conclusion: This study suggests a decrease in brain pH in patients with MDD, potentially associated with recurrent episodes and longer illness duration. As suggested from previous animal model studies, altered brain energy metabolism, leading to decreased pH, may serve as a potential transdiagnostic endophenotype for MDD and other neuropsychiatric disorders.

Fig. 1. Workflow of the selection of postmortem brain pH datasets.

The studies that reported matching brain tissue pH between patient and control groups and/or excluding samples with a pH below a certain threshold were considered as pH-balanced studies.

Fig. 2. Decreased pH in the postmortem brains of patients with MDD.

Forest plot of meta-analysis comparing postmortem brain pH between patients with MDD and control individuals. 95% CI, 95% confidence interval; SE, standard error; SMD, standardized mean difference.

Fig. 3. Decreased brain pH in female patients associated with recurrent episodes and illness duration in MDD.

A Scatter plot of z-score-transformed pH values based on sex and diagnosis. The z-scores were calculated within each dataset. Two-way ANOVA: effect of sex: p = 0.0007; effect of diagnosis: p < 0.0001; interaction: p = 0.043. Red bars indicate mean values for each group. The p-values from Sidak’s multiple comparison test following the two-way ANOVA are shown. B–D Effect of recurrence of depression episodes on brain pH levels by sex. Scatter plot of z-score-transformed pH values of control subjects, patients with MDD with a single episode, and those with recurrent episodes of depression in the female datasets (B), male datasets (C), and combined female and male datasets (D). A total of 535 samples from seven datasets were analyzed. The z-scores were calculated within each dataset. The p-values from Tukey’s multiple comparison test following the one-way ANOVA are shown. E Scatter plot of the duration of illness by sex. A total of 114 samples from five datasets were analyzed. The p-value from the unpaired t-test is shown. CON, control subjects; MDD, major depressive disorder.

Fig. 4. Negative correlation between brain pH and duration of illness.

Forest plot of meta-analysis of correlation between brain pH and duration of illness in patients with MDD. 95% CI, 95% confidence interval; CC, correlation coefficient; SE, standard error.