Umeå University's logo

Emotional unstable personality disorder (EUPD; previously borderline personality disorder, BPD) is associated with excess natural-cause mortality, comorbid medical conditions, poor health habits and stress related epigenomic alterations. Previous studies demonstrated that GrimAge – a state-of-the-art epigenetic age (EA) estimator – strongly predicts mortality risk and physiological dysregulation. Herein, we utilize the GrimAge algorithm to investigate whether women with EUPD and a history of recent suicide attempts exhibit EA acceleration (EAA) in comparison to healthy controls. Genome-wide methylation patterns were measured using the Illumina Infinum Methylation Epic BeadChip in whole blood from 97 EUPD patients and 32 healthy controls. The control group was significantly older (p < 0.0001) and reported lesser exposure to violent behavior in both youth and adulthood (p < 0.0001). Groups were otherwise comparable regarding gender, BMI, or tobacco usage (p > 0.05). EA estimator DNAmGrimAge exceeded chronological age by 8.8 and 2.3 years in the EUPD and control group, respectively. Similarly, EAA marker AgeAccelGrim was substantially higher in EUPD subjects when compared to controls, in both univariate and multivariate analyzes (p < 0.00001). Tobacco usage conferred substantial within-group effects on the EA-chronological age difference, i.e., 10.74 years (SD = 4.19) compared to 6.00 years (SD = 3.10) in the non-user EUPD group (p < 0.00001). Notably, past alcohol and substance abuse, use of psychotropic medications, global assessment of functioning, self-reported exposure to violent behavior in youth and adulthood, later completed suicide (N = 8) and age at first suicide attempt did not predict EAA in the EUPD group (p > 0.05). These results underscore the importance of addressing medical health conditions along with low-cost preventative interventions aimed at improving somatic health outcomes in EUPD, such as efforts to support cessation of tobacco use. The independency of GrimAge to other EA algorithms in this group of severely impaired EUPD patients, suggest it may have unique characteristics to evaluate risk of adverse health outcomes in context of psychiatric disorders.