| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Department Group of Clinical Medicine, University of Oslo; Department of Anesthesiology, Rikshospitalet University Hospital; and Department of Plastic and ENT Surgery, Colosseum Clinic, Oslo, Norway
Address correspondence and reprint requests to Luis Romundstad, University of Oslo, Department Group of Clinical Medicine, Department of Anesthesiology, Rikshospitalet University Hospital, N-0027 Oslo, Norway. Address e-mail to luisro{at}medisin.uio.no.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Breast augmentation with subglandular silicone implantation is a highly standardized procedure in a homogenous population of young adult women and gives substantial initial inflammatory and pressure-related pain (2). Hence breast augmentation surgery is an appropriate clinical model for studying the analgesic effect of antiinflammatory analgesic drugs on postoperative pain (3). Glucocorticoids have potent antiinflammatory properties (4). They have analgesic effects after oral, orthopedic, and laparoscopic surgery (5–7). After hip arthroplasty, a single dose of methylprednisolone 125 mg IV caused analgesia lasting 24 h and opioid-sparing effects for 72 h (7). Their antiemetic properties in addition to the analgesic effects make them particularly useful (8,9). After ambulatory surgery, glucocorticoids also increase satisfaction scores and quality of recovery facilitating discharge (10). The glucocorticoids reduce prostaglandin synthesis by reducing activity of phospholipase A2, and by selectively blocking the expression of cyclooxygenase (COX)-2 mRNA but have a negligible effect on COX-1 (4). They also inhibit other mediators of inflammatory hyperalgesia such as tumor necrosis factor-
, interleukin 1ß, and interleukin 6 (11). Perioperative single dose glucocorticoid, even in a large dose (i.e., methylprednisolone 30 mg/kg), were not associated with any adverse effects in a meta-analysis with more than 1900 patients included (12).
The COX-2 specific inhibitors give similar postoperative analgesia as the nonspecific COX-1 and COX-2 inhibitors (nonsteroidal antiinflammatory drugs) (13). COX-2-specific inhibitors do not impair platelet function (14). They reduce the risk for bleeding and gastrointestinal adverse effects compared with the nonselective nonsteroidal antiinflammatory drugs (14,15). However, the selective COX-2 inhibitors have a number of other possible side effects such as renal and heart failure, hypertension, thrombotic cardiovascular events, and gastrointestinal bleeding (16–18). A recent trial investigating the safety of parecoxib and valdecoxib treatment during a 10-day period after coronary bypass surgery concluded that these drugs were associated with an increased incidence of myocardial infarction, cardiac arrest, stroke, and pulmonary embolism (17). Despite extensive perioperative use of COX-2 inhibitors, we have not been able to find any reports on serious renal or cardiovascular complications associated with its short-term use after noncardiac surgery on healthy patients. In this setting, these side effects are probably not a concern.
We hypothesized that a single presurgical dose of methylprednisolone would have analgesic and antiemetic effects after augmentation mammoplasty. Therefore, the aims of the present study were to compare the analgesic and antiemetic effects of a single dose of methylprednisolone and a single dose of the COX-2-specific inhibitor parecoxib with placebo. Because of its known analgesic efficacy from previous postoperative trials (19), parecoxib was used as a comparator drug to monitor study sensitivity.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
We follow the CONSORT recommendations for reporting randomized, controlled clinical trials (21) (Fig. 1). The Regional Committee for Research Ethics in Norway and The Norwegian Medicines Agency approved the study protocol. Patients gave their written consent to participate after verbal and written information. They were informed that they would receive either an active analgesic drug or saline as a placebo drug before surgery. After surgery, oral acetaminophen 500 mg with codeine 30 mg, one to two tablets orally, was permitted as rescue analgesic when the patients had so much pain that they felt an analgesic was required. When they left the clinic, the patients were given acetaminophen with codeine tablets for 5 days (up to two tablets four times per day). If they required more than eight tablets a day, they were told to call the clinic or the doctor responsible for the study (LR) for more effective rescue pain relief and exclusion of any surgical complications.
|
A person not involved in the treatment and follow-up of patients randomized the patients in blocks of nine to 1 of 3 groups of equal size using a list of random numbers, according to the Moses-Oakford algorithm (22). Block size and randomization code were not revealed to the investigators until all measurements and calculations had been entered into the database.
Healthy women between 20 and 45 years old, scheduled for breast augmentation surgery with subglandular silicone implants, were candidates for inclusion. Final inclusion was decided in the morning just before surgery. Patients were excluded if they had known sensitivity to any of the test drugs, including sulfur allergy, dyspepsia, history of peptic ulcer, bleeding tendency, renal failure, serious or symptomatic heart, lung, or any other organ disease, drug abuse, or chronic analgesic medication.
Before the start of surgery, the patients were sedated, and the local anesthetic was injected. The sedation was deep enough to bring almost all the women to sleep during most of the procedure and consisted of an IV propofol infusion 5–10 mg · kg–1 · h–1 (Diprivan®;AstraZeneca, Södertälje, Sweden) plus IV fentanyl 0.1–0.2 mg (Leptanal®; JansenCilag, Bersee, Belgium). During sedation, the women were monitored with pulse oximetry, electrocardiogram, and arterial blood pressure. Lidocaine 10 mg/mL with adrenaline 5 µg/mL (Xylocaine with adrenaline®; AstraZeneca), 80 mL diluted in 200 mL of saline, was injected in the spatium between the pectoralis muscle and the mammary gland. This dose (roughly 14 mg/kg) seems to be large, but data from Rygnestad et al. (23) indicate that a dose of 20 mg/kg of lidocaine with adrenaline is safe in breast augmentation. However, the same precautions as for major regional anesthesia were taken to deal with toxicity.
The same type of surgery, i.e., breast augmentation with subglandular silicone implants, was performed on all patients. Surgical access was either submammary or subareolar (Table 1). The same two surgeons (KS and HR) did all surgery on all patients.
|
Methylprednisolone (Solu-Medrol®; Pfizer Inc., New York, New York) 125 mg, parecoxib (Dynastat®; Pfizer Inc.) 40 mg, and placebo (NaCl, 9 mg/mL) were prepared at Rikshospitalet University Hospital by a doctor not in contact with the observers or patients. Test drugs were diluted with saline to fill a 10-mL syringe, marked with patient number and the possible test drugs, and appeared identical for all persons involved in the trial. After transport to the day-case clinic, the coded test drug was given IV during 2 min by the anesthesiologist before the sedation of the patient was started.
The test and rescue drugs were provided by the Department of Anesthesiology, Rikshospitalet University Hospital and the Colosseum Clinic, Oslo, Norway. No other grants were received.
The surgeons interviewed, informed, and included the patients during the consultation some days before surgery. In the morning before surgery, they repeated the information and instructed the patients how to fill out the questionnaire rating their pain intensity after surgery.
All pain evaluations were done both at rest and while performing a circular movement with the arms towards the midline and pressing the hands together using the pectoralis major muscles (dynamic pain). Present pain intensity was evaluated with an 11-point Numeric Rating Scale (0–10 NRS; 0 = no pain and 10 = unbearable pain).
On the day of surgery, pain intensity was assessed 1 h after the end of anesthesia (103 ± 9 min [mean ± sd] after test-drug administration) with the help of a physician or nurse in the postanesthesia care unit. Before the patients left the hospital 120 min after end of surgery, pain intensity was assessed, and the nurse or physician made sure that the patients understood how to fill out the rest of the questionnaire. Then the patients left the clinic to go to the patients' hotel or home. Subsequently, pain intensity was assessed at 4 and 6 h after the end of anesthesia. For patients requiring rescue analgesic before 6 h, data were computed in two different ways:
Pain at bedtime (resting and dynamic), worst pain at 0–24 h, and average pain at 0–24 h were assessed during the first 24 h after end of surgery using the NRS. Three patients in the methylprednisolone group, 4 in the parecoxib group, and 12 in the placebo group required rescue analgesia before 1 h after the end of anesthesia. They scored pain intensity before being given rescue and were not excluded.
The primary measure of efficacy was the summed pain intensity from 1 to 6 h, calculated as the sum of time-weighted pain intensity scores (using the 0–10 NRS) from 1–6 h (using LOCF) as an area under the curve (24). Secondary outcomes were time to rescue analgesic, rescue usage for each day during the 6-day study period, a composite pain-intensity-rescue analgesic score, nausea, vomiting, and fatigue.
In trials where rescue analgesics are given early to a significant proportion of study patients, the numbers of "true" pain observations after test drugs decline rapidly, and the common method of LOCF will soon be based on very few observations and therefore have reduced validity (25). Therefore, we calculated a composite score based on actual pain observations (both before and after rescue, as opposed to LOCF) and rescue analgesic consumption during the first 6 h. This resulted in two composite variables: one calculated from actual pain at rest and rescue usage and one calculated from actual dynamic pain and rescue usage. We modified the Silverman et al. (26) method for calculating a composite score as follows: (a) Rank all subjects (n = 204) according to their sum of actual pain intensities observed during the first 6 h (area under the curve), (b) express the difference of each treated subject's rank from the mean rank as a percentage of the mean rank, (c) perform the same steps for rescue analgesic use, and (d) for each subject, the sum of the two percentage differences calculated for actual pain rank and rescue analgesic usage rank constitutes the composite score.
During the first 24 h, all side effects, with special attention to postoperative nausea and vomiting (PONV), fatigue, feeling of general illness, and dizziness, were assessed at all observation time points by an open question: "Do you experience any of the mentioned or any other side effects?" If present, the type and severity (no = 0, mild = 1, moderate = 2, or severe = 3) of the side effect were noted. Spontaneously reported side effects were assessed in the same way. We also graded PONV from 0 to 3, where 0 = no nausea, 1 = mild nausea, 2 = moderate nausea, and 3 = severe nausea with vomiting.
On the subsequent 5 days after surgery, pain intensity at rest and dynamic pain were measured in the morning (before the first dose of rescue analgesic) and at bedtime. Worst pain, average pain during the day and night, and to what extent the pain affected sleep and activities of daily life were also assessed each day throughout the study period using the 0–10 NRS.
After the first 24 h, the patients were told to report all spontaneously occurring adverse effects and complications. If present, the type and severity were described.
When the patients arrived home or in the patients' hotel, a clinic nurse or doctor phoned and reminded them to continue filling out the pain intensities and side effects on the questionnaire. They were also phoned on the first and fifth day after surgery. On the fifth day, they were reminded to finish the questionnaire and send it back to the clinic.
Sample size was calculated based on results from 12 pilot patients. Mean pain intensity of 4 (0–10 NRS) in the placebo group from 1 to 6 h and a common sd of 2 were assumed. A pain reduction (summed pain intensity at 1–6 h) of at least 30% was considered clinically important. Given these assumptions, it was calculated that a sample size in each of the three groups of 54, in a one-way analysis of variance (ANOVA), would have 90% power to detect at the 0.05 level a difference in means of this magnitude (characterized by a variance of means of 0.320) (nQuery Advisor 5.0; Statistical Solutions, Saugus, Massachusetts). Taking the uncertain compliance of an outpatient population into consideration, we decided to increase the number of included patients to at least 70 in each group to allow a dropout rate of 20%.
Demographic data were analyzed by one-way ANOVA. Adverse effects (categorical data) were analyzed with Fisher's exact test. Analyzing pain intensity variables and rescue analgesic consumption were performed using the one-way ANOVA with Bonferroni correction in the post hoc multiple comparisons. To compare time to rescue among the groups, Kruskal-Wallis and subsequent Mann-Whitney U-test were performed. Significance was determined at the P < 0.05 level. Data are presented as mean ± 95% confidence interval (CI) or mean ± sd. The SPSS statistical program, version 12.0 (SPSS Inc., Chicago, Illinois) was used for analyses of all data.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The three groups had similar distribution of age, weight, and height. All patients had similar duration of surgery and anesthesia (Table 1).
Both active drugs decreased pain at rest and dynamic pain intensity from 1–6 h after surgery compared with placebo (summed pain intensity [LOCF]: methylprednisolone versus placebo; P < 0.03; parecoxib versus placebo; P < 0.001; summed dynamic pain intensity [LOCF]: methylprednisolone versus placebo; P < 0.01; parecoxib versus placebo; P < 0.001) (Fig. 2A).
|
Mean time (hours) to the first rescue analgesic dose was significantly longer in both the methylprednisolone (3.4; 95% CI, 2.5–4.3) and the parecoxib group (4.2; 95% CI, 3.0–5.4) compared with placebo (2.2; 95% CI, 1.9–2.5) (methylprednisolone, P < 0.05; parecoxib, P < 0.01; Fig. 3A). The active-drug groups also consumed significantly fewer rescue analgesics compared with the placebo group during the first 6 h (Table 2). After 6 h, there were no significant differences among the groups in consumption of rescue analgesics.
|
|
Because of the frequent early re-medication rate with rescue analgesics (Fig. 3), we calculated composite scores from actual pain observations and rescue analgesic consumption during the first 6 h. This composite variable of the active drugs was similar, and both were significantly superior to that of placebo for pain at rest, as well as dynamic pain (P < 0.001) (Fig. 2B).
There were no significant differences between methylprednisolone and parecoxib in any pain intensity variables during the study (Fig. 2) and no significant differences in any efficacy variables between the two types of surgical access (submammary or subareolar).
Methylprednisolone was significantly superior to both parecoxib (P < 0.05) and placebo (P < 0.001) regarding the total numbers of adverse events (of any type) during the first 24 h (Fig. 4A). After methylprednisolone, but not parecoxib, there was a significantly reduced incidence and severity of PONV compared with placebo (P < 0.001; Fig. 4B). Methylprednisolone, but not parecoxib, produced significantly less fatigue than placebo (P < 0.05; Fig. 4C). Side effects such as dizziness and a general feeling of illness were relatively common but did not differ significantly among the groups. Parecoxib was not significantly different from placebo for any side effect.
|
From the day after surgery and until the end of the fifth day after surgery, we found no significant differences among the groups concerning sleep disturbed by pain, pain at rest, and dynamic pain in the morning or at bedtime, average pain during the day and night, worst pain, effects on activities of daily life, and analgesic consumption or adverse effects.
Two patients developed a postoperative hematoma (one in the placebo group, and one in the parecoxib group) during the first 48 h. One patient in the placebo group developed a postoperative infection during the first 2 wk.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This was a single-center study with a large number of patients of the same sex and within narrow age limits, undergoing a highly standardized surgical procedure by the same two surgeons (KS and HR), and producing considerable pain and inflammatory reaction from the surgical tissue trauma. Satisfactory assay sensitivity was documented in the present trial by a significant difference in outcome measures between the active control drug (parecoxib) and placebo.
About half of the patients in all groups had taken rescue analgesics during the first two hours, partly reflecting the nursing staff routine of offering these ambulatory patients analgesics (Fig. 3). This secured pain relief before the patients left the clinic after approximately 120 minutes after the end of surgery. However, a problem in the interpretation of the pain intensity data based on the conventional LOCF method was created. The number of patients who had not taken rescue analgesia decreased rapidly, and the statistical calculations were therefore based on a decreasing number of real pain observations during the first six hours, reducing the validity of the statistical calculations based on LOCF (25). We therefore calculated a composite score from actual pain observations and rescue analgesic consumption during the first six hours. This composite score (26) was similar for the two active drugs, and both were significantly superior to placebo for pain at rest and dynamic pain (Fig. 2, A and B).
Thus, in the present trial, the data based on LOCF and the data based on integration of actual pain and rescue analgesics consumed led to the same conclusion, namely, both active drugs were significantly better than placebo. The integrated assessment differentiated even better between the active drugs and placebo.
Both the actual pain intensity (raw pain data) and the use of rescue analgesics were significantly less in the active-drug groups compared with the placebo group (Fig. 2B). The active drugs were, in fact, documented to be active analgesic drugs!
Some studies on, e.g., major abdominal surgery (5) have not been able to document the definite analgesic effect of glucocorticoids. Confounding factors, such as differences in anesthesia methods, surgical technique, other adjuvant therapies, main focus on other outcome variables than pain, and small-sized studies with low power, may have contributed to negative outcome in those studies. Trials using less than dexamethasone 8 mg or methylprednisolone 40 mg seem to be less likely to give quantifiable analgesic effects (5). During and after major surgery, the endogenous glucocorticoids reach such high levels that the benefit of additional glucocorticoids on pain may be small. But even in major surgery, benefits have been demonstrated on other variables such as reduced inflammatory response, improved pulmonary function, less fatigue, increased appetite, less PONV, and more rapid convalescence (6,27).
In the present study, methylprednisolone reduced the total number of adverse events, most importantly PONV, during the first 24 hours. PONV is the main cause of delay in discharge home and unanticipated hospital readmission after ambulatory surgery (28). After breast surgery, the incidence of PONV can be 60%–80% (29). In this trial, the incidence was 60% in the placebo group. After methylprednisolone, PONV was effectively reduced both in incidence (30%) and severity during the first 24 hours compared with placebo (Fig. 4B). Parecoxib did not reduce PONV compared with placebo. These findings are in agreement with those of Apfel et al (30).
In a systematic review, Rubin and Hotopf (31) suggested that glucocorticoids may attenuate fatigue immediately after operations. Our trial confirmed a reduced incidence and severity of postoperative fatigue after methylprednisolone administration compared with placebo (Fig. 4C).
In a previous study, we found a reduced opioid consumption lasting at least 72 hours after a single dose of methylprednisolone 125 mg given the morning after major orthopedic surgery (6). Less pain intensity in all groups after 24 hours may be the reason that such a sustained pain-relieving effect of methylprednisolone could not be detected in the present trial.
In conclusion, methylprednisolone 125 mg IV given before breast augmentation surgery had analgesic and rescue analgesic-sparing effects significantly superior to placebo and comparable to those of parecoxib 40 mg IV. Methylprednisolone, but not parecoxib, reduced postoperative emesis and fatigue. These effects of methylprednisolone are particularly useful in short-stay surgical patients.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Jokela, J. Ahonen, M. Tallgren, M. Haanpaa, and K. Korttila Premedication with pregabalin 75 or 150 mg with ibuprofen to control pain after day-case gynaecological laparoscopic surgery Br. J. Anaesth., June 1, 2008; 100(6): 834 - 840. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Kardash, F. Sarrazin, M. J. Tessler, and A. M. Velly Single-Dose Dexamethasone Reduces Dynamic Pain After Total Hip Arthroplasty Anesth. Analg., April 1, 2008; 106(4): 1253 - 1257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kjetil, T. K. Sem, S. Ellen, and R. Johan The Prolonged Postoperative Analgesic Effect When Dexamethasone Is Added to a Nonsteroidal Antiinflammatory Drug (Rofecoxib) Before Breast Surgery Anesth. Analg., August 1, 2007; 105(2): 481 - 486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. F. White, H. Kehlet, J. M. Neal, T. Schricker, D. B. Carr, F. Carli, and the Fast-Track Surgery Study Group The Role of the Anesthesiologist in Fast-Track Surgery: From Multimodal Analgesia to Perioperative Medical Care Anesth. Analg., June 1, 2007; 104(6): 1380 - 1396. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Jokela, J. Ahonen, M. Valjus, T. Seppala, and K. Korttila Premedication with controlled-release oxycodone does not improve management of postoperative pain after day-case gynaecological laparoscopic surgery Br. J. Anaesth., February 1, 2007; 98(2): 255 - 260. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
