Stress and Pain. Predictive (Neuro)Pattern Identification for Chronic Back Pain: A Longitudinal Observational Study

Abstract

Introduction Low back pain (LBP) leads to considerable impairment of quality of life worldwide and is often accompanied by psychosomatic symptoms. Objectives First, to assess the association between stress and chronic low back pain (CLBP) and its simultaneous appearance with fatigue and depression as a symptom triad. Second, to identify the most predictive stress-related pattern set for CLBP for a 1-year diagnosis. Methods In a 1-year observational study with four measurement points, a total of 140 volunteers (aged 18–45 years with intermittent pain) were recruited. The primary outcomes were pain [characteristic pain intensity (CPI), subjective pain disability (DISS)], fatigue, and depressive mood. Stress was assessed as chronic stress, perceived stress, effort reward imbalance, life events, and physiological markers [allostatic load index (ALI), hair cortisol concentration (HCC)]. Multiple linear regression models and selection procedures for model shrinkage and variable selection (least absolute shrinkage and selection operator) were applied. Prediction accuracy was calculated by root mean squared error (RMSE) and receiver-operating characteristic curves. Results There were 110 participants completed the baseline assessments (28.2 ± 7.5 years, 38.1% female), including HCC, and a further of 46 participants agreed to ALI laboratory measurements. Different stress types were associated with LBP, CLBP, fatigue, and depressive mood and its joint occurrence as a symptom triad at baseline; mainly social-related stress types were of relevance. Work-related stress, such as “excessive demands at work”[ b = 0.51 (95%CI -0.23, 1.25), p = 0.18] played a role for upcoming chronic pain disability. “Social overload” [ b = 0.45 (95%CI -0.06, 0.96), p = 0.080] and “over-commitment at work” [ b = 0.28 (95%CI -0.39, 0.95), p = 0.42] were associated with an upcoming depressive mood within 1-year. Finally, seven psychometric (CPI: RMSE = 12.63; DISS: RMSE = 9.81) and five biomarkers (CPI: RMSE = 12.21; DISS: RMSE = 8.94) could be derived as the most predictive pattern set for a 1-year prediction of CLBP. The biomarker set showed an apparent area under the curve of 0.88 for CPI and 0.99 for DISS. Conclusion Stress disrupts allostasis and favors the development of chronic pain, fatigue, and depression and the emergence of a “hypocortisolemic symptom triad,” whereby the social-related stressors play a significant role. For translational medicine, a predictive pattern set could be derived which enables to diagnose the individuals at higher risk for the upcoming pain disorders and can be used in practice.

Topics & Concepts