| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Department of Anesthesiology, University of Washington, Seattle, Washington; Global Outcomes Research, Pharmacia, Skokie, Illinois; and The Mountain-Whisper-Light Statistical Consulting, Seattle, Washington
Address correspondence to D. J. Pavlin, Department of Anesthesiology, University of Washington, 1959 N.E. Pacific, Seattle, WA 98195. Address e-mail to jpavlin{at}u.washington.edu Reprints will not be available.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
7, and 24% were delayed in Phase 1 recovery by pain. Pain scores were lower if local anesthetic or ketorolac was administered intraoperatively (22% and 26% respectively). Fentanyl dose during recovery correlated with maximum pain scores; fentanyl dose was 42% less if ketorolac was administered intraoperatively. In females, the recovery fentanyl dose increased in proportion to the intraoperative fentanyl dose. The maximum pain score was predictive of total recovery time (135, 172, and 212 min of recovery for maximum pain scores of 0–3, 4–6, and 7–10, respectively; P < 0.001). We conclude that improvements in pain therapy are warranted to improve patient comfort and to expedite recovery. IMPLICATIONS: Moderate to severe pain is common after ambulatory surgery and is a frequent cause of delayed discharge. Postoperative pain, opioid-related side effects, and time to discharge were less when nonsteroidal antiinflammatory drugs or local anesthetics were used intraoperatively to prevent pain before patient awakening.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Ideally, we would like to optimize pain control after surgery to enhance patient comfort and expedite recovery. This requires an understanding of the determinants of postoperative pain and a knowledge of how pain and analgesic requirements relate to recovery time. Accordingly, this study was designed as a prospective survey of pain after ambulatory surgery. The specific goals of the study were to determine pain severity, analgesic use, and the relationship of pain and analgesic therapy to recovery time before discharge. A secondary goal was to identify factors predictive of these outcomes. In this observational study, we investigated the effects of patient age, history of analgesic use, type and duration of anesthesia and surgery, and intraoperative use of local anesthetics and analgesics as possible predictors of postoperative outcomes.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Data obtained from patients included age, height, weight, history of surgery, recent analgesic use (past 7 days), analgesic side effects, history of smoking, postoperative nausea/vomiting (PONV), or motion sickness. The type and duration of surgery and the type and dose of anesthetic drugs, analgesics, and antiemetics administered as part of anesthesia were recorded. Times, doses, and routes of administration of analgesics, antiemetics, and other drugs used during recovery were recorded until the time of discharge. A variety of oral analgesics were used. An equivalent dose was calculated for each drug to permit summing equivalent doses where one equivalent dose was aspirin, 650 mg; ibuprofen, 200 mg; acetaminophen, 650 mg; naproxen, 250 mg; codeine, 60 mg; oxycodone, 6.7 mg; and hydrocodone, 7.5 mg. The equivalent doses used were obtained from the medical literature (9–12). Given that there is some variability in the results of comparative studies, these dose equivalencies must be viewed as approximate.
Numeric pain scores (0–10 where 0 = no pain and 10 = worst imaginable pain) were recorded at 15-min intervals during recovery until discharge. Duration of pain scores >3 and total duration of analgesic therapy were determined for each patient. Duration of analgesic therapy was defined as the time from first to last analgesic dose plus 20 min. Twenty minutes was added because nurses routinely waited 20 min after an analgesic dose (either IV or oral) before discharging a patient. Other milestones of recovery recorded included time to transfer from Phase 1 to Phase 2 recovery and time to actual discharge. The data are presented as mean values (±SE) for each group. All information was obtained prospectively by two study coordinators who observed patients throughout their hospital course.
When recovery was delayed (>50 min in Phase 1 or >70 min in Phase 2), nurses caring for the patient indicated up to three causes of delay, prioritized in order of importance as described in a previous report (1). Causes were selected from a preprinted list that included medical, surgical, and system factors but also permitted unique entries.
No attempt was made to alter normal practice. Anesthesia was administered by residents in training or nurse anesthetists supervised by an attending anesthesiologist. Anesthesia personnel used standard postoperative order forms to order IV pain and antiemetic medications. Oral pain medications were ordered by surgeons as part of patients’ discharge medications. Recovery room nurses administered analgesics as they considered appropriate; they used a numeric pain scale of 0–10 to assess pain until the patient attained a score of 
3 or said they were comfortable, which is the usual practice in our institution. Antiemetics were administered when patients complained of nausea or vomited.
Descriptive data are presented as means (±SE) or percentages. Means were compared between categories of patients by using unpaired Student’s t-tests or analysis of variance. Percentages for dichotomous and categorical variables were compared by using 
2 or Fisher’s exact tests, as appropriate.
Multivariate linear regression analysis was used to assess the independent, unique significance (the R2 value, expressed as a percentage) of various factors as predictors of four major outcomes: maximum pain, fentanyl use during recovery, duration of analgesic therapy, and total recovery duration. The details of the analyses are described in Appendix 1.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
|
Of factors possibly related to PONV, 70% received one or more prophylactic antiemetics before or during surgery, 25% had a history of PONV, 38% had a history of motion sickness, and 91% were nonsmokers (13). Pain scores obtained at serial time intervals for the first 90 min of recovery are shown in Figure 1, demonstrating the average score and the change over time for different surgical procedures. After 90 min, the number of patients in each group diminished considerably because of patient discharge. The characteristics of the in-hospital recovery period are shown in Table 2. Maximum pain, duration of pain therapy, total recovery time, and the incidence of PONV were numerically greater after hernia, laparoscopic, and plastic surgery. Overall, 37% of patients required no analgesics.
|
|
In Table 3, the results of uni- and multivariate analyses are shown for those factors found to be uniquely predictive of the four primary end points of the study. Maximum pain was most strongly related to the type of surgery. Infiltration of local anesthetic in the wound or ketorolac administered during surgery correlated with a one-point reduction in pain score (22%–26% reduction in pain score, respectively). Factors that correlated significantly with increased pain by univariate analysis, but not when adjusted for other factors, included younger patient age, longer duration of surgery (or anesthesia), larger mean end-tidal sevoflurane concentration, and larger doses of fentanyl per kilogram body weight administered pre- and intraoperatively.
|
|
Variation in total recovery duration was independently predicted by the presence or absence of PONV (R2 = 8%; P 0.001), by maximum pain score (R2 = 7%; P 0.001), and by the combined effects of surgical procedure, intraoperative fentanyl dosage, and the interactions between the two (R2 = 13%; P = 0.002). Thus, recovery was 202 vs 142 min in patients with or without PONV, respectively; 135 min in patients with maximum pain scores of 0–3; 173 min with pain scores of 4–6; and 212 min with pain scores of 7–10. Use of ketorolac was associated with a 25-min shorter duration of recovery (P = 0.03).
Because PONV was an important predictor of recovery duration, we computed odds ratios for various potential predictors of PONV (Table 4). Intraoperative use of ketorolac was predictive of a reduced probability of PONV (odds ratio, 0.45; 95% confidence interval, 0.23–0.90). Increasing the fentanyl dose—either the pre- or intraoperative dose or the cumulative dose—was associated with an increased likelihood of PONV (odds ratio, 1.4; 95% confidence interval, 1.16–1.69 for cumulative fentanyl dosage). Univariate analysis and logistic regression were also used to investigate the relationship between various factors possibly predictive of an increased likelihood of PONV. Univariate analysis suggested a significant association between PONV and pre- or intraoperative fentanyl dosage, type of surgery, surgery duration, and use of ketorolac or local anesthetic intraoperatively. However, only the combined effects of type of surgery, cumulative fentanyl dose, and an interaction term were independently predictive of PONV by logistic regression when corrected for other factors. The latter accounted for 17% of variability in the likelihood of PONV.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
7 on a 10-point scale). Pain was the most common cause of Phase 1 recovery delays, affecting 24% of patients overall. Pain was a significant independent predictor of total recovery duration, and patients who did require analgesics were discharged 54 minutes later than patients who did not require analgesic therapy. Of the various factors evaluated as predictors of pain severity, the type of surgery performed was the most important; pain was most severe after hernia, laparoscopy, and plastic surgery. However, use of a local anesthetic by the surgeon or the administration of ketorolac intraoperatively was associated with a one-point reduction in pain score(22%–26% decrease in numeric pain score, respectively). Local anesthetic infiltration was not as effective as one might have expected, possibly because surgeons were less motivated to attain adequate anesthesia in the presence of a general anesthetic or because anesthetized patients were unable to provide feedback regarding its effectiveness.
Despite several reports attesting to the benefits of preoperative or intraoperative administration of NSAIDs, most patients in this study did not receive a NSAID before the end of surgery (with the exception of patients having knee arthroscopy) (14,15). In our population, IV ketorolac at the time of wound closure was associated with less severe pain (pain score, 3.1 vs 4.17; P = 0.006), diminished fentanyl requirements in the recovery room (0.29 vs 0.68 µg/kg; P = 0.05), a shorter recovery duration (147 vs 172 minutes; P = 0.03), and a lesser likelihood of PONV (odds ratio, 0.45; 95% confidence interval, 0.23–0.9), illustrating the utility of NSAIDs before completion of surgery. Conceivably, preoperative use of a NSAID might have been even more effective. Reubin et al. (16) recently reported that patients had less severe postoperative pain if they received an oral cyclooxygenase 2 inhibitor before, as opposed to after, knee arthroscopy. Traditional NSAIDs are sometimes considered undesirable in surgical patients because of potential adverse effects on platelet function and hemostasis (17). This was relevant to plastic surgery patients in our study. The cyclooxygenase 2 inhibitors are a newer class of NSAID that decrease formation of mediators of pain but do not impair platelet function (18). Overall, it appears that NSAIDs were underused in this study, given the salutary effects on pain, recovery time, and postoperative emesis observed in this and other studies.
In this study, we also observed that the dose of fentanyl administered in recovery was positively correlated with the amount of fentanyl received during surgery in the four operations performed in women under general anesthesia. This observation was somewhat unexpected, because one might anticipate that larger doses of opioids (fentanyl) intraoperatively would reduce opioid requirements after surgery. There are several possible explanations for this finding. The first is that female patients who received more fentanyl intraoperatively in response to blood pressure and heart rate changes during surgery were more needy because of random variation in sensitivity to opioids or physiologic responses to noxious stimuli. A second possibility is that surgical procedures were more invasive in women who required more fentanyl. A third possibility is that repeated dosing with fentanyl intraoperatively induced an acute state of tolerance, hyperalgesia, or both. Hyperalgesia has been demonstrated in rats after repeated small doses of fentanyl (19). Chia et al. (20) recently reported increased postoperative pain and fentanyl requirements after abdominal hysterectomy in women who received large-dose fentanyl (15 µg/kg) versus small-dose fentanyl (1 µg/kg) intraoperatively. Clarification of the exact nature of the relationship between intraoperative opioid use and analgesic requirements after ambulatory surgery would require further investigation in a larger population composed of both sexes.
Of note, increasing the dose of fentanyl intraoperatively was also predictive of an increased incidence of PONV and more prolonged recovery. Overall, PONV occurred in 34% of patients and, when present, delayed discharge by approximately one hour. This occurred despite the fact that 70% of patients received one or more antiemetics prophylactically before the end of surgery.
This study may be criticized because it was performed in a single university institution. The results may therefore not be applicable to all other institutions. For that reason, the details of anesthetic administration and analgesic therapy have been included in detail to enable the reader to judge how patient care in this study might relate to other institutions. The observational nature of the study also does not permit drawing definitive conclusions regarding cause-and-effect relationships of associations. It is also possible that the regulated periodicity of obtaining pain scores (which were readily available to nurses caring for patients) may have altered the usual speed of treating pain. However, because nurses routinely used the same scoring system at similar intervals, it is unlikely that the conduct of the study had a very significant effect on outcomes.
Overall, this study provides basic information regarding the severity of pain after ambulatory surgery, the dose of analgesics required to control pain, and the significance of pain as a determinant of recovery duration, and it delineates factors that contribute to variability in these outcomes. The data demonstrate that NSAIDs and local anesthetic techniques reduce postoperative pain, use of opioids, and opioid-related side effects. The data also suggest that at least in women, intraoperative fentanyl may have undesirable dose-related effects on pain and recovery duration. Finally, the study demonstrates that improvements in speed of recovery that might be anticipated by improvements in analgesic therapy may be dependent on simultaneous prevention of emetic episodes.
Appendix 1
Multivariate linear regression was used to assess the unique significance of various factors as predictors of the four major outcomes of the study: maximum pain, fentanyl use during recovery (µg/kg), duration of analgesic therapy, and total recovery duration. Univariate analyses were first performed with Pearson’s correlation coefficient or linear regression to identify variables that were significantly related to each outcome. Factors entered into the preliminary analysis included patient age, weight, sex, smoking history, history of PONV, history of recent opioid or NSAID use (in the past week), surgery duration, anesthesia duration, surgery type, pre-/intraoperative fentanyl dose (µg/kg), use of local anesthesia by the surgeon, administration of ketorolac during surgery, mean end-tidal sevoflurane concentration, and, where appropriate, maximum pain and PONV. An R2 was then calculated to describe the proportion of variation of an outcome that could be explained by a given variable considered alone (i.e., by univariate analysis, expressed as a percentage in tables). We then attempted to identify the unique contribution (R2 unique) of each variable to predicting outcome. The latter was defined as the proportion of variation of an outcome that could be related uniquely to a given variable when other factors were controlled for. For a continuous variable, the unique contribution of a predictor variable was calculated as the change in R2 from linear regression when the predictor variable was added into a reference model including other statistically significant predictors. The statistical significance of the change in R2 was based on the F test. For a given outcome variable, the reference model included all statistically significant predictors except when two predictor variables had a strong correlation (r > 0.5), in which case only one of the variables was used in the reference model. For example, duration of surgery and duration of anesthesia were highly correlated (r = 0.98); therefore, only duration of surgery was used in the reference model. Because lesser correlations persisted between some predictor variables, values of R2 unique may underestimate the true maximum effect of a given predictor variable. In fact, the true effect may lie somewhere between the R2 calculated by univariate analysis (i.e., uncorrected) and the R2 unique (i.e., corrected). In a parallel fashion, the unique contribution of each variable in predicting the dichotomous outcome of nausea or vomiting in the recovery period was calculated by comparing logistic regression models with and without the given variable. Statistical significance was calculated from the change in log likelihood, and a descriptive pseudo-R2 was calculated on the basis of the log likelihood. Odds ratios showing the change in risk of nausea or vomiting are also presented for the logistic regression models. An odds ratio of 1.0 indicates that the risk of nausea or vomiting does not vary across levels of a predictor variable. In these analyses, the R2 unique represents the minimum unique contribution of a particular factor to the end point when corrected for other potentially significant variables.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. F. White Pain management after ambulatory surgery - Where is the disconnect?/La prise en charge de la douleur apres une chirurgie ambulatoire : ou est-ce que ca coince ? Can J Anesth, April 1, 2008; 55(4): 201 - 207. [Full Text] [PDF]
|
|
|
|
 |
|
 T. I. Usichenko, S. Kuchling, T. Witstruck, D. Pavlovic, M. Zach, A. Hofer, H. Merk, C. Lehmann, and M. Wendt Auricular acupuncture for pain relief after ambulatory knee surgery: a randomized trial Can. Med. Assoc. J., January 16, 2007; 176(2): 179 - 183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. T. Awad and F. Chung Factors affecting recovery and discharge following ambulatory surgery: [Les facteurs influencant la recuperation et la sortie apres une operation en chirurgie ambulatoire]. Can J Anesth, September 1, 2006; 53(9): 858 - 872. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S. Liu, W. M. Strodtbeck, J. M. Richman, and C. L. Wu A Comparison of Regional Versus General Anesthesia for Ambulatory Anesthesia: A Meta-Analysis of Randomized Controlled Trials Anesth. Analg., December 1, 2005; 101(6): 1634 - 1642. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Mattila, J. Toivonen, L. Janhunen, P. H. Rosenberg, and M. Hynynen Postdischarge Symptoms After Ambulatory Surgery: First-Week Incidence, Intensity, and Risk Factors Anesth. Analg., December 1, 2005; 101(6): 1643 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. B. Gordon, J. L. Dahl, C. Miaskowski, B. McCarberg, K. H. Todd, J. A. Paice, A. G. Lipman, M. Bookbinder, S. H. Sanders, D. C. Turk, et al. American Pain Society Recommendations for Improving the Quality of Acute and Cancer Pain Management: American Pain Society Quality of Care Task Force Arch Intern Med, July 25, 2005; 165(14): 1574 - 1580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. F. White Update on ambulatory anesthesia Can J Anesth, June 1, 2005; 52(suppl_1): R10 - R10. [Full Text] [PDF]
|
|
|
|
|
|
H. Ma, J. Tang, P. F. White, A. Zaentz, R. H. Wender, A. Sloninsky, R. Naruse, R. Kariger, R. Quon, D. Wood, et al. Perioperative Rofecoxib Improves Early Recovery After Outpatient Herniorrhaphy Anesth. Analg., April 1, 2004; 98(4): 970 - 975. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Pavlin, K. D. Horvath, E. G. Pavlin, and K. Sima Preincisional Treatment to Prevent Pain After Ambulatory Hernia Surgery Anesth. Analg., December 1, 2003; 97(6): 1627 - 1632. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
