///2017 Abstract Details
2017 Abstract Details2019-08-02T15:54:53-06:00

Comparison of the Apfel risk score to a newly derived risk score for PONV following cesarean delivery under spinal anesthesia

Abstract Number: SAT-36
Abstract Type: Original Research

Rebecca Anderson M.D.1 ; Rebecca Anderson MD2; Ronald B George MD, FRCPC3; Mary Cooter MS4; Ashraf S Habib MBBCh, MSc, MHSc, FRCA5


There are limited data about risk factors for postoperative nausea and vomiting (PONV) in obstetric patients. The commonly used simplified Apfel risk score includes female gender, and expected receipt of postoperative opioids, which might limit its applicability to obstetric patients receiving neuraxial opioids, since all will have at least 2 risk factors. We therefore performed this study to derive a risk score for PONV in women undergoing cesarean delivery (CD) and to compare its performance to that of Apfel’s.


We included data from two multicenter randomized controlled trials investigating PONV. Anesthetic management in both studies included spinal anesthesia with 12 mg hyperbaric bupivacaine, 15 mcg fentanyl and 150 mcg morphine. We only included patients who did not receive prophylactic antiemetics. Potential risk factors for PONV based on literature with some added pregnancy specific factors were collected (table). We identified factors significantly associated with PONV in our cohort using multivariate mixed models and derived a new simplified risk score by assigning 1 point to each identified risk factor The association of this score and Apfel’s with PONV was assessed using Cochran-Armitage Trend tests. We compared our score to the Apfel score using Mann-Whitney U test of difference in AUC, and Integrated Discrimination Improvement (IDI).


Data from 260 patients were included in the analysis. Of the risk factors evaluated, only smoking pattern and history of PONV after CD or history of morning sickness were significantly associated with PONV (table). We therefore created a score that gives 1 point to history of PONV following CD or morning sickness, and 1 point for stopping smoking during pregnancy plus 1 point for never smoking. There was a significant association between PONV and higher values of our score (risk of PONV with 0, 1, 2, 3 points of 33, 39, 52 and 65%, p= 0.002), but not the Apfel score (risk of PONV with 2, 3, 4 points of 39, 57, 61%, p=0.12). The estimated AUC of our score and that of Apfel’s were 0.63 [0.56,0.70] and 0.58 [0.52,0.65] (p= 0.20). The IDI was 0.032, meaning the average PONV risk prediction improved by 3.2% using our score compared to Apfel’s.


There was no statistically significant difference in the performance of our derived risk score compared to the Apfel score. However, our score showed modest incremental improvements by using variables specific to CD patients.

SOAP 2017