Investigating patterns of association between analgesic consumption and reported pain scoring in adult cardiothoracic surgical patients

Abstract

Background: Cardiac surgery is invasive resulting in high pain intensity which can be difficult to control. Pain is commonly assessed and treated in the ICU/HDU with reported pain scores (RPS). Analgesic consumption (AC) has been suggested as a proxy or supplement to RPS giving more objective measures. This study examined these patterns of association in the context of physiological, clinical, and demographic data. Methods: A retrospective study of 261 cardiac patients in the ICU/HDU. The analysis had 3 phases, multiple linear regression, logistic regression, and principal component analysis (PCA) which included 56 explanatory variables and 2 outcomes, AUC of pain scores (AUC-RPST) and a binary classification of pain control. The exploration of the associations between applied PCA to understand the latent structures. Results: The final model explained 57.8% of the variance in AUC-RPST (AIC=3348, p<2.2e-16). Higher total opioids and NSAIDs were consistently associated with pain intensity and variability and a lower likelihood of pain control. PCA revealed components associated with higher variance in cH+, HCO3, CVP, history of CKD, AF, CRD were most strongly associated with higher AUC-RPST followed by variance in electrolytes, walking, chest-tubes, and SYSBP. Procedure type and components associated with bypass time, nicotine, marching, and other aspects of magnesium and lactate were negatively associated with AUC-RPST. Conclusions: Healthcare providers should be aware that higher total amounts of analgesics do not necessarily lead to pain control. Patients who have renal insufficiency, hemodynamic instability, altered metabolism, and higher variation in electrolytes have higher AUC-RPST. The findings suggest that while analgesics explain variability of pain outcomes, multiple physiological, demographic, and clinical factors improved understanding of these associations revealing a complex web.