| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Department of Anesthesiology at the University of Utah Medical Center, Salt Lake City, Utah
Address correspondence to Jennifer J. Davis, MD, Department of Anesthesiology, University of Utah Medical Center, 30 North 1900 East, Room 3C 444, Salt Lake City, UT 84132. Address e-mail to jennifer.davis{at}hsc.utah.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
When opioids are used as the primary form of analgesia, it is imperative to define dose-response relationships for clinical end-points such as analgesia and respiratory depression. This is especially important in patients receiving long-term opioid treatment because doses associated with respiratory depression and analgesia may be several times larger than those for opioid-naive patients (2). During the postoperative period, patients who have been consuming opioids chronically are often treated, ineffectively, with conventional doses of opioid. This may be followed by escalating doses of opioids in conjunction with sedatives or anxiolytics. Even with careful monitoring, this practice can be associated with serious adverse events.
Most clinicians are justifiably concerned about administering unusually large doses of opioid in the less adequately monitored setting of the hospital ward. Although it is true that patients who chronically consume opioids are less susceptible to some side effects of opioids, such as pruritus and nausea, excessive sedation and respiratory depression may be more common in this population than previously appreciated (2). Considering the possibility of excessive sedation in this population and their propensity to report higher pain scores (2), safe and effective opioid delivery can be challenging.
We hypothesized that a preoperative fentanyl challenge in conjunction with pharmacokinetic simulation could be used to identify the fentanyl effect-site concentration (Ce) associated with respiratory depression. These data could then be used to optimize opioid delivery during the intraoperative and postoperative periods for each individual while avoiding respiratory depression. A secondary aim of our study was to compare the preoperative fentanyl challenge with the preoperative review of opioid consumption (i.e., oral use of hydrocodone, oxycodone, and morphine) standardized in terms of morphine equivalents in terms of their relationship with the fentanyl Ce at steady-state PCA settings 24 h after surgery. We hypothesized that the fentanyl challenge would provide a better correlation than the preoperative assessment of morphine equivalents with postoperative fentanyl Ce levels.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
All patients were instructed to stop opioid medications after midnight on the evening before surgery and no preoperative sedation was used. Before the induction of anesthesia, 100% oxygen was administered through a well-sealed mask while monitoring the electrocardiogram, noninvasive arterial blood pressure, pulse oximetry, and capnography. An IV infusion of fentanyl, 2 µg ·kg–1 · min–1 (based on ideal body weight), was started. No adjunctive drugs were administered during the fentanyl infusion and no tactile or verbal stimulation was allowed. The fentanyl infusion was continued until the patient demonstrated depression of spontaneous ventilation (defined as a respiratory rate of <5 breaths per minute as measured by capnography). At this point, the fentanyl infusion was stopped and general anesthesia was induced with propofol 1.5–2.0 mg/kg. Anesthesia was maintained using desflurane in air and oxygen. The duration of the fentanyl infusion was recorded for each patient.
Using pharmacokinetic simulation software (STANPUMP, Stanford, CA; this software is freely available from Steve L. Shafer at URL http://anesthesia.stanford.edu/pkpd/), the fentanyl Ce near the onset of respiratory depression was estimated for each patient using pharmacokinetic variables described by Shafer et al. (4). Figure 1 presents a simulation of the fentanyl challenge. Figure 2 presents a series of simulated fentanyl infusions that was used to determine the hourly intraoperative and postoperative infusion rate. This infusion rate was selected to provide a fentanyl Ce that was 30% of that associated with respiratory depression for each individual. This threshold was selected based on existing data for fentanyl, which suggest that concentrations producing analgesia are approximately 30% of those associated with respiratory depression (5,6). In an additional simulation, we estimated the fentanyl Ce during the preoperative and postoperative course based on the dosing of fentanyl for a representative patient (see Appendix).
|
|
The target analgesic infusion rate of fentanyl was maintained intraoperatively and into the recovery room. Postoperatively, to provide additional safety, only 50% of the predicted hourly analgesic requirement was administered as a basal infusion; the remainder was available to the patient as demand PCA doses. The lockout interval was set to 15 min, which is longer than traditionally used (6). This interval was selected to ensure that the peak effect of each dose would be evident before the patient was able to administer a subsequent dose.
At 4-h intervals, the basal infusion rate was increased or decreased to maintain a demand dose rate of 2–3 doses per hour. For example, in patients using <1 interval dose per hour, the basal infusion was decreased by 20%. Alternatively, for those patients using more than 3 doses per hour, the basal infusion was increased by 20%. The basal infusion was stopped for any patient with a respiratory rate <10 breaths/min. Each patient received oxygen by nasal cannula and was monitored with continuous pulse oximetry. In addition, an investigator was assigned to monitor each patient continuously for the first 24 h after surgery.
After patients reached near steady-state PCA settings (defined as the PCA in use for at least 24 h and no PCA adjustments for at least 8 h), plasma fentanyl concentrations and arterial pH, Paco2, and Pao2 were measured. All plasma fentanyl concentrations were sampled from arterial blood 30 min after the most recent PCA demand dose. At steady-state, plasma fentanyl concentrations were used to approximate fentanyl Ce. Samples were analyzed using radioimmunoassay. The assay calibration points were from 0.2 to 6.0 ng/mL. The estimated error occurring within the standard curve was 15%.
Demographic data were described as means ± sd. Preoperative morphine equivalents, steady-state postoperative fentanyl concentrations, and blood gas values were reported as median and interquartile ranges. Using simple linear regression, an exploration was made of the relationship between morphine equivalents and plasma fentanyl concentrations at steady-state PCA settings and the relationship between the predicted fentanyl Ce near the onset of respiratory depression and plasma fentanyl concentrations at steady-state PCA settings. On the basis of the strong relationships observed in this linear regression analysis, additional analysis was performed to explore whether morphine equivalents or predicted fentanyl Ce near the onset of respiratory depression were a better predictor of postoperative plasma fentanyl levels at steady-state PCA settings. To accomplish this, multiple linear regression was used to model postoperative plasma fentanyl concentration at steady-state infusion (the response variable) as a function of the two explanatory variables: predicted fentanyl Ce at the time of respiratory depression and preoperative morphine equivalents.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
|
No patient required stimulation to breathe, naloxone, or any major intervention related to excessive sedation during the postoperative course. The highest Paco2 observed in any patient was 48 mm Hg. The median Paco2 was 41 mm Hg with an interquartile range of 39–46 mm Hg. The median pH was 7.42 with an interquartile range of 7.39–7.43. The median Pao2 was 72 with an interquartile range of 64–80 mm Hg.
Figure 3 depicts a pharmacokinetic simulation of the fentanyl Ce that resulted from the preoperative fentanyl challenge, the intraoperative continuous infusion, and the postoperative basal infusion and PCA dosing for a representative patient in this study. The calculations for this sample patient are shown in the Appendix. In this example, the patient required 11 min to achieve a respiratory rate <5 breaths per minute during the fentanyl challenge. The intraoperative fentanyl infusion was continued until patient admission to the recovery room. Subsequently, the basal infusion and PCA were initiated.
|
The number of adjustments in the basal infusion rate based on PCA usage is presented in Table 2. In seven patients, no adjustment was required from initial PCA settings. Eight patients required a single increase in the basal infusion. Only one patient required a decrease in the basal infusion predicted by the preoperative response to fentanyl. For three patients, more than one adjustment in PCA settings was required.
|
Simple linear regression examining the relationship between morphine equivalents and plasma fentanyl concentrations at steady-state PCA settings revealed a coefficient of determination (r2) of 0.61 and the relationship between the predicted fentanyl Ce near the onset of respiratory depression and plasma fentanyl concentrations at steady-state PCA settings revealed an r2 value of 0.91. Multiple linear regression found a statistically significant coefficient for the regression of fentanyl concentration at steady-state PCA settings on predicted Ce associated with respiratory depression (coefficient = 0.327, P < 0.00001). The coefficient for the regression of fentanyl concentration at steady-state PCA settings on preoperative morphine equivalents was near zero and not statistically significant (coefficient = 0.082, P = 0.81). Morphine equivalents were deleted from the regression and the model was rerun with simple linear regression (Fig. 4). The multiple regression model fit was not impaired by eliminating morphine equivalents (P = 0.81).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
The intent of this work was to 1) implement a fentanyl challenge using a brief large-dose continuous fentanyl infusion to estimate intraoperative and postoperative opioid requirements just before surgery, 2) use this information to guide the administration of fentanyl during the intraoperative and postoperative phases of patient care, and 3) evaluate this technique in terms of its impact on respiratory depression during the postoperative phase of analgesic therapy. We selected fentanyl for use in this study because of its rapid onset, ease of titration, lack of active metabolites, well-defined pharmacokinetic and pharmacodynamic profile, and widespread use among health care providers. Our study hypothesis was that a preoperative fentanyl challenge in conjunction with pharmacokinetic simulations to estimate the fentanyl Ce associated with respiratory depression could be used to individualize effective postoperative analgesic dosing regimens in chronically opioid-consuming patients. Metrics used to evaluate this hypothesis were a review of PCA usage (i.e., number of demand interval doses per hour) over the immediate 24-hour postoperative course as a measure of meeting opioid requirements for each patient and an arterial blood gas evaluation to assess the presence of respiratory acidosis after achieving steady-state PCA settings. The results of this study confirmed our study hypothesis. The most important findings of this study were that 1) using a fentanyl challenge before surgery led to implementation of PCA settings in combination with a continuous basal infusion that did not require significant adjustment and 2) arterial blood gas analysis revealed no evidence of respiratory depression. Of the 12 patients who required an adjustment in PCA settings, all but one required a modest increase in their basal infusion rate; no patient required stimulation to breathe, naloxone, or any other intervention related to excessive sedation.
As reported by Rapp et al. (2), chronically opioid-consuming patients typically report higher pain scores than opioid-naïve populations. With this in mind, we made no attempt to achieve a specific pain score during the postoperative period. Instead, the basal infusion rate was increased or decreased to the point where the patient would use 2–3 demand doses per hour. This demand dose rate suggested that additional analgesia was still available to the patient yet confirmed that the basal infusion was not excessive. With this approach, primary consideration was given to careful monitoring of oxygenation, respiratory rate, and appropriate use of the PCA.
An important consideration in this patient population is that the development of tolerance to various opioid effects may occur at different rates. For example, tolerance to respiratory depression may occur more slowly than analgesia or euphoria (9), suggesting that escalating doses of opioids may not provide improved analgesia yet may suppress respiration. Although existing data suggest that analgesia occurs at a threshold that is approximately 30% of the drug level required to produce respiratory depression (5,6), there may be wide individual variation in this relationship. Indeed, for some chronically opioid-consuming patients, the ratio of drug levels that produce analgesia versus respiratory depression may be dangerously high, explaining the paradox of increased sedation and respiratory depression in these patients (2). This underscores the potential danger of using pain scores as the singular end-point for postoperative opioid dosing.
A secondary aim of our study was to compare how well the preoperative fentanyl challenge versus the preoperative assessment of morphine equivalents was at correlating with the fentanyl Ce at steady-state PCA settings. In our multiple regression model analysis, we found that the fentanyl challenge outperformed the preoperative assessment of morphine equivalents. This would suggest that because of pharmacokinetic and possible pharmacodynamic differences between opioids, as well as covariates such as age and weight that impact opioid pharmacokinetic behavior, preoperative opioid use, standardized in terms of equianalgesic morphine equivalents, is not as effective in predicting postoperative analgesic requirements as the fentanyl challenge.
We do not know what the intraoperative and postoperative fentanyl requirements would have been in these patients had they not received the initial loading dose of fentanyl during the fentanyl challenge. We speculate that they would have been somewhat larger. However, because this technique requires a fentanyl challenge to safely predict a target analgesic Ce, we do not feel that this would add to the clinical relevance of the study.
As part of the fentanyl challenge, we used pharmacokinetic modeling software to estimate the fentanyl Ce near the onset of respiratory depression. This estimation has several limitations that merit discussion. First, when using a continuous large-dose fentanyl infusion of brief duration, simulations revealed that the fentanyl Ce initially increases rapidly over the first 6 minutes and then continues to increase, although at a slower rate (Fig. 1). It is important to recognize that an error in measuring the duration of the continuous infusion or poor technique in measuring the respiratory rate can lead to underestimation or overestimation of the fentanyl Ce associated with respiratory depression. For example, an error of plus or minus 1 minute during a 10-minute infusion results in a ±10% difference in the fentanyl Ce.
Second, we used simulations of continuous fentanyl infusions to target fentanyl Ce values that would provide effective analgesia. One important feature of fentanyl pharmacokinetics is that during prolonged continuous infusions, the fentanyl concentration continues to slowly increase. In instances where a set continuous infusion is used over several days, the actual fentanyl Ce may exceed the target fentanyl Ce for analgesia. This was one of the compelling reasons we set the basal infusion of fentanyl that accompanied the fentanyl PCA to only 50% of the infusion rate required to reach the target fentanyl Ce for analgesia. This required the patient to self-administer the remaining fentanyl required to achieve 30% of the threshold in the fentanyl Ce for respiratory depression. It is important to note that no patient enrolled in this study demonstrated respiratory depression during the study period.
Third, we used simulations of fentanyl Ce as opposed to fentanyl plasma concentrations to determine thresholds for respiratory depression and analgesia. We did this to account for the hysteresis between plasma concentrations and the concentration at the site of drug action (the Ce) during rapid changes in fentanyl administration (i.e., the initial phase of continuous infusions, bolus administration, or termination of an infusion). Based on this feature of fentanyl pharmacokinetic behavior, we did not measure a plasma fentanyl level near the onset of respiratory depression because of the disequilibrium between the plasma and effect-site fentanyl concentrations. We did, however, obtain a sample for fentanyl assay once the basal infusion and PCA settings had been in use for at least 24 hours and had remained unchanged for more than 8 hours. At this point, we assumed that the plasma fentanyl concentration had reached equilibrium with the fentanyl Ce. Our results confirmed this assumption.
An additional limitation of this study was that none of the patients enrolled in this study suffered from respiratory impairments such as sleep apnea or chronic obstructive pulmonary disease. It is not clear how respiratory disorders such as these may affect the postoperative course in patients who receive large-dose opioid infusions.
In summary, we evaluated the use of a preoperative fentanyl challenge in chronically opioid-consuming patients to estimate the amount of fentanyl required to meet analgesic requirements after a surgical procedure associated with substantial postoperative pain. To accomplish this goal we used pharmacokinetic-based simulations to individualize the estimates of postoperative analgesic requirements. We found that this approach allowed us to fine tune the postoperative administration of fentanyl to a patient population known for their difficulty with regard to optimizing pain control and who are at risk for excessive sedation when, as is often required with this patient population, doses exceed conventional dosing recommendations. Although this technique has been a valuable application of pharmacokinetic modeling to individualize care, this line of investigation merits additional work to further define its efficacy and minimize the risk of adverse events.
|
Appendix: Example of a "Fentanyl Challenge" in a Typical Study Patient |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
2. A continuous fentanyl infusion is initiated at a rate of 2 µg · kg–1 · min–1 (170 µg/min). This is continued until the onset of respiratory depression (respiratory rate < 5 breaths/min as measured by capnography).
3. Respiratory depression occurred after 11 min of the fentanyl infusion. Using Figure 1, the effect-site concentration (Ce) is estimated as 24 ng/mL.
4. Analgesia is achieved at a fentanyl Ce that is 30% of the fentanyl Ce associated with respiratory depression. The target analgesic Ce is 0.30 x 24 ng/mL or 7.2 ng/mL.
5. Using Figure 2, the total hourly infusion rate to achieve a fentanyl Ce of 7.2 ng/mL is approximately 6 µg · kg–1 · h–1.
6. An intraoperative fentanyl infusion of 6 µg · kg–1 · h–1 (approximately 510 µg/h) is maintained for the duration of the surgery and into the postanesthesia care unit (approximately 5.5 h).
7. Postoperative patient-controlled analgesia (PCA) settings: Fifty percent of the total hourly analgesic requirement is administered as a basal infusion; 510 µg/h x 0.5 = 255 µg/h. The remaining 50% of the hourly analgesic requirement is available to the patient in divided demand doses. Lockout of 15 min = 4 demand doses per hour. 255 µg/4 = approximately 63 µg/dose or 1.3 mL/dose.
Initial PCA settings: basal rate = 5.1 mL/h, lockout interval = 15 min, demand dose = 1.3 mL.
8. The basal rate is adjusted according to the average demand dose utilization. For patients using < 1 demand dose per hour, the basal rate is decreased by 20%. For patients using >3 demand doses per hour, the basal rate is increased by 20%.
*Respiratory rate and sedation should be carefully assessed before any adjustment.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: Example of a...References |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
