Can remifentanil use in obstetrics be improved by optimal patient-controlled analgesia bolus timing?

BACKGROUND The safety of patient-controlled i.v. analgesia (PCA) with remifentanil for obstetrical analgesia remains a matter of concern. The efficacy of remifentanil bolus application, that is, the coincidence between pain and remifentanil effect-site concentration, may be improved by forecasting contractions, but it is not known whether such a technique would also improve safety. METHODS We recorded pain intensity during labour continuously using a handheld dynamometer in 43 parturients. Using these data, we compared different models in their ability to predict future contractions. In addition, we modelled remifentanil effect-site concentration using three simulated modes of bolus administration, with and without prediction of future contractions. RESULTS The average duration of pain during contractions recorded by the dynamometer was 45 [14 standard deviation (sd)] s. The time interval between painful contractions was highly variable, with a mean of 151 (31 sd) s during the first and 154 (52 sd) s during the second recording. Using a simple algorithm (three-point moving average), the sd of the difference between predicted and observed inter-contraction intervals can be reduced from 0.95 to 0.79 min. However, the coincidence between remifentanil concentration and pain during contraction is not substantially improved when using these models to guide remifentanil bolus application. CONCLUSIONS Because of the large variability of inter-contraction intervals, the use of prediction models will not influence the mean remifentanil concentration in-between contractions. Using models predicting future contractions to improve the timing of remifentanil PCA bolus administration will not diminish the need of continuous clinical surveillance and other safety measures.