The performance of the heart rate variability‐derived Newborn Infant Parasympathetic Evaluation Index as a measure of early postoperative pain and discomfort in infants—A prospective observational study
Laura M. Verweij, Jaap T. S. Kivits, Frank Weber
Abstract
Abstract Background The heart rate variability‐derived Newborn Infant Parasympathetic Evaluation (NIPE™) Index is a continuous noninvasive tool for the assessment of pain and discomfort in infants. Little is known about its performance in the early postoperative setting, where assessment of pain/discomfort is usually based on discontinuous observational scoring systems or personal experience of medical staff. Aims To investigate the performance of the NIPE as a measure of early postoperative pain and/or discomfort in infants. Methods The potential of the NIPE to detect pain/discomfort, as assessed by two clinical scoring systems (FLACC and COMFORT‐B scale), was investigated in postoperative infants (0–2 years). Results Receiver operating curve (ROC) analyses investigating the power of the NIPE to distinguish between comfort and pain/discomfort, revealed areas under the curve (AUC) of 0.77 for the FLACC, 0.81 for the COMFORT‐B score, and 0.77 for a combination of FLACC & COMFORT‐B. Logistic regression analysis provided initial evidence that the NIPE is an independent predictor of a FLACC score ≥4 and/or a COMFORT‐B score ≥17, though R 2 values were below .2. NIPE values associated with a FLACC ≥4 (48 [45–56]), a COMFORT‐B score ≥17 (47 [42–53]), and a FLACC ≥4 & COMFORT‐B ≥17 (47 [42–57]) were lower than NIPE values associated with a FLACC <4 (60 [53–68], 95% CI of difference −14 to −8, p < .0001), a COMFORT‐B score <17 (61 [54–68], 95% CI of difference −16 to −10, p < .0001), and a FLACC <4 & COMFORT‐B score <17 (60 [53–68], 95% CI of difference −15 to −8, p < .0001). We found no evidence of a predictive value of the NIPE regarding the occurrence of pain. Conclusions The NIPE detected pain and discomfort in infants after general anesthesia with reasonable areas under the ROC curve (±0.8), whereas it was not predictive of clinically detectable pain or discomfort.