1Department of Anaesthesia, McMaster University, Hamilton, Canada. 2Department of Anaesthesia, Vrije Universiteit, Brussels, Belgium. 3Department of Obstetrics and Gynaecology, University of Glasgow, Glasgow, UK. 4Department of Surgical Gastroenterology, University of Copenhagen, Hvidovre, Denmark. 5Department of Medical Statistics, London School of Hygiene and Tropical Medicine, London, UK. 6Department of Anaesthesia, Hôpital Tenon, Paris, France. 7Department of Primary Care and Population Sciences and Department of Social Medicine, University of Bristol, Bristol, UK. 8Department of Haematology, University of Barcelona, Barcelona, Spain. 9Department of Anaesthesia, Universität Mainz, Mainz, Germany. 10Department of Pharmacology, University of Verona, Verona, Italy. 11Department of Medicine, Birmingham University, Birmingham, UK. 12Department of Gastroenterology, L.Sacco University Hospital, Milan, Italy. 13Department of Haematology, University Hôpital-Dieu de Paris, Paris, France. 14Ckdt Communications Ltd, Allschwil, Switzerland*Corresponding author: Department of Anaesthesia, HSC 2U4, McMaster University, Hamilton, Canada L8N 3Z5
Declaration of interest. This study was funded by F. Hoffmann-La Roche Ltd.
Accepted for publication: October 12, 2001
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Methods. This prospective, randomized multicentre trial evaluated the risks of death, increased surgical site bleeding, gastrointestinal bleeding, acute renal failure, and allergic reactions, with ketorolac vs diclofenac or ketoprofen administered according to their approved parenteral and oral dose and duration of treatment. Patients were followed for 30 days after surgery.
Results. A total of 11 245 patients completed the trial at 49 European hospitals. Of these, 5634 patients received ketorolac and 5611 patients received one of the comparators. 155 patients (1.38%) had a serious adverse outcome, with 19 deaths (0.17%), 117 patients with surgical site bleeding (1.04%), 12 patients with allergic reactions (0.12%), 10 patients with acute renal failure (0.09%), and four patients with gastrointestinal bleeding (0.04%). There were no differences between ketorolac and ketoprofen or diclofenac. Postoperative anticoagulants increased the risk of surgical site bleeding equally with ketorolac (odds ratio=2.65, 95% CI=1.514.67) and the comparators (odds ratio=3.58, 95% CI=1.936.70). Other risk factors for serious adverse outcomes were age, ASA score, and some types of surgery (plastic/ear, nose and throat, gynaecology, and urology).
Conclusion. We conclude that ketorolac is as safe as ketoprofen and diclofenac for the treatment of pain after major surgery.
Br J Anaesth 2002; 88: 22733
Keywords: analgesics non-opioid, diclofenac; analgesics non-opioid, ketoprofen; analgesics non-opioid, ketorolac; complications
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Study population
Adults over 18 yr old, undergoing elective major surgery were entered in the trial. The study procedure defined certain exclusions (patients with known sensitivity to any of the study drugs or other NSAIDs, patients in whom NSAIDs were contraindicated, who were pregnant or lactating, who were to undergo minor, emergency or day-case surgery, or who were ASA physical status class V).
For the purpose of the study, major surgery was defined as a procedure of such complexity as to require admission to hospital for more than 24 h and to require injections of NSAID for relief of postoperative pain. The surgical operation performed was documented and then classified for analysis as orthopaedic (e.g. joint replacement, osteotomy, discotomy); abdominal (e.g. colectomy, small bowel resection, open cholecystecytomy); gynaecological (e.g. total hysterectomy, myometrectomy, oophorectomy); urological (e.g. nephrectomy, suprapubic prostatectomy); plastic/ear, nose and throat (ENT) (e.g. extensive skin graft, breast reconstruction, rhinoplasty, faciomaxillary procedures); and others (including cardiac, vascular, and thoracic surgery). Arthroscopic, laparoscopic, and endoscopic procedures were not included.
Study design
The study was a prospective, randomized multicentre trial of ketorolac vs diclofenac or ketoprofen to evaluate the risks of five primary serious adverse outcomes; these were death, increased surgical site bleeding, gastrointestinal bleeding, acute renal failure, and severe allergic reactions. They were defined according to standard diagnostic criteria (Appendix 1). Patients were followed for 30 days after their surgery. Randomization was carried out in blocks at each centre to ensure a balanced allocation between ketorolac and the centres designated comparator. All patients who were randomized and for whom we had information about outcomes were included in the analysis (intention-to-treat).
Each of the study drugs was administered according to the approved product label for each drug. The maximum daily dose and duration of treatments were: ketorolac, parenteral 90 mg day1 for 2 days followed by oral 40 mg day1 for up to 7 days; diclofenac, parenteral 150 mg day1 for 2 days followed by oral 150 mg day1 for up to 7 days; and ketoprofen, parenteral 200 mg day1 for 2 days followed by oral 200 mg day1 for up to 7 days. If additional analgesia was required, an opioid could be used. Abstracted data were subjected to random audit and tested for inter-observer reliability.
Data analysis
Each of the serious adverse outcomes was analysed by the MantelHaenszel test or the Fisher exact test for rare outcomes to test the null hypothesis. Likelihood ratio tests were used to examine heterogeneity of outcomes in different subsets of patients. The MannWhitney test was used for continuous variables, and the chi-squared test, or Fisher exact test was used for categorical variables. The only stratification used was by study centre. Risk was assessed by multiple stepwise logistic regressions. Data management was performed by Quintiles Scotland Ltd, and the statistical analysis was performed at the London School of Hygiene and Tropical Medicine. The Steering group of principal investigators was responsible for the design, conduct, and interpretation of this study and was blinded until the final analysis was completed. The Study Safety Committee was un-blinded and provided regular status reports to the UK Medicines Control Agency.
Sample size
An estimate of the required sample size was made using data from a retrospective post-marketing surveillance study which reported a rate of 3.6% for a composite outcome (patients who had died, had increased surgical bleeding, or had gastrointestinal bleeding) in a cohort of patients who received ketorolac and opioid.7 Using this rate, a sample of 16 000 was calculated to provide over 90% power to detect a relative risk of 1.3 for this composite outcome. Interim power calculations were planned after recruitment of 10, 30, and 60% to determine when a power of at least 80% was achieved to detect a relative risk of 1.5 for the composite outcome. The power calculation was based on a one-sided test using the normal approximation to the Poisson with a 0.025 level of significance, equal to 0.05 on a two-sided test. The significance levels for stopping rules that were established a priori by the Safety Committee for death was P<0.01 with a relative risk of 3.0, and for serious adverse outcomes was P<0.001 with a relative risk of 3.0 separately for increased surgical site bleeding, gastrointestinal bleeding, acute renal failure, and severe allergic reactions. The Safety Committee recommended to stop the study after reviewing the data on the estimated 60% of patients. The time to obtain agreement from the CPMP, and the 30-day observation period resulted in additional patients being entered in the study before the database was closed. A total of 11 302 patients were randomized. This sample provided over 80% power to detect a relative risk of 1.5 at the rate of 1.24%, which was found for the composite outcome.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
The most frequent outcome was increased surgical site bleeding in 117 patients (1.04%) of whom 61 received ketorolac, and 56 received one of the comparators. The risk of surgical site bleeding was 3.05 times higher with the use of postoperative anticoagulants. Forty-one per cent of patients received an anticoagulant after surgery including fractionated low-molecular-weight heparin (LMWH) in 2929 patients (26%), and low-dose unfractionated heparin (UFH) in 1552 patients (13.8%). The risk of surgical site bleeding was generally higher with low-dose UFH than with fractionated LMWH. More patients received anticoagulants after orthopaedic (55%), gynaecology (55%), abdominal (47%), and urology (34%) surgery than after plastic/ENT surgery (14%). Patients who received ketorolac (odds ratio=2.65, 95% CI=1.514.67) or comparator NSAID (odds ratio=3.58, 95% CI=1.936.70) were equally likely to experience a higher risk of surgical site bleeding when an anticoagulant was administered (Table 3).
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Gastrointestinal effects
Gastrointestinal perforations, ulceration, and bleeding (PUBs) have been reported with all NSAIDs after chronic use.8 9 However, the risk of serious gastrointestinal adverse effects with NSAIDs, including ketorolac, is very low after short-term perioperative use10. A review of 15 postoperative placebo-controlled trials reported only one case of serious gastrointestinal bleeding in 1520 patients, of whom 927 had received an NSAID.10 Several studies have reported an increased risk of serious gastrointestinal bleeding in patients over 60 yr old, in women, in patients with a history of peptic ulceration, in smokers, and in patients who consume excessive amounts of alcohol.1113 A retrospective post-marketing surveillance study in over 20 000 patients who received ketorolac or opioid,7 reported a small increased risk of gastrointestinal bleeding with ketorolac (odds ratio=1.30, 95% CI=1.111.52) but no significant increased risk of surgical bleeding (odds ratio=1.02, 95% CI=0.951.10). These risks were significantly increased further in patients over 70 yr, and in patients who received excessive doses of ketorolac for more than 5 days. In our study, there were only four patients (0.04%) who had gastrointestinal bleeding. None of these patients received ketorolac.
Postoperative surgical site bleeding
The use of NSAIDs after surgery may increase the risk of bleeding by their antiplatelet effects;2 however, the evidence of increased bleeding is conflicting. Some studies have reported no increased bleeding with ketorolac,1 or diclofenac14 when administered after transurethral prostatectomy, a procedure that is commonly associated with postoperative bleeding. In contrast, aspirin or other NSAIDs taken preoperatively increased the risk of postoperative bleeding after prostatectomy.15 Bleeding time is increased with NSAID therapy and for ketorolac this increase is 312468 s16 (normal 600 s). However, the antiplatelet effects of NSAIDs do not appear to significantly alter postoperative coagulation and may be protective for myocardial infarct and thromboembolism.17
In our study, the risk of postoperative surgical site bleeding was increased when anticoagulant drugs were used for thromboprophylaxis after surgery, but was no different between ketorolac and the two other NSAIDs. Low-dose UFH was associated with a higher risk of bleeding (2.5%) than fractionated LMWH preparations (1.3%). A recent meta-analysis of randomized trials in general surgery has reported that low-dose LMWH was safer than low-dose UFH, but at higher doses there was an increased risk of major haemorrhage with both preparations.18 A possible interaction between NSAID and anticoagulant drugs cannot be ruled out, but this does not appear to involve a clinically important increase in bleeding time.19 Some types of surgery in our study were associated with a higher risk of surgical site bleeding, independent of the use of postoperative anticoagulants. Plastic/ENT surgery increased the risk of surgical site bleeding by 3.5 times, gynaecological surgery increased the risk by 2.7 times, and urological surgery increased the risk by 2.5 times. A recent study in 215 patients after breast reconstruction found no increased incidence of haematoma with ketorolac,20 and in our study we found no difference for increased surgical site bleeding between ketorolac and the two comparators in patients who had undergone plastic/ENT surgery or other types of surgery. Whereas the absolute risk of increased surgical site bleeding appears to be small, further studies are needed to evaluate the potential interaction and increased risk with concomitant NSAID and anticoagulant use in specific types of surgery.
Renal effects
Cyclo-oxygenase-1 and -2 are important components of the control of glomerular filtration and excretion of urine.21 NSAIDs can interfere with renal function through the inhibition of both forms of cyclo-oxygenase. There are four toxic effects of NSAIDs on the kidney: acute ischaemic renal insufficiency, acute interstitial nephritis, analgesic-associated nephropathy, and progressive hypertensive nephropathy. Reports of acute renal failure after ketorolac and other NSAIDs are mainly the result of acute ischaemic effects.22 Other factors that increase the risk of acute renal failure include hypovolaemia, cardiac failure, ascites, and liver cirrhosis.21 A large retrospective cohort found no significant association with ketorolac in patients who developed postoperative acute renal failure.23 A meta-analysis of eight randomized trials of NSAID effects on postoperative renal function reported only a minor transient decrease in creatinine clearance after ketorolac, diclofenac, indomethacin, or ibuprofen.24 Ten patients (0.1%) had acute renal failure in our study, three cases were reported after ketorolac and seven cases after the comparator NSAIDs. There was no significantly increased risk of acute renal failure in 42 patients with a history of renal insufficiency, or 72 patients with a history of congestive heart failure.
Allergic reactions
Allergic reactions to NSAIDs are uncommon and include skin rash, bronchospasm, and anaphylaxis. It has been estimated that between 5 and 10% of asthmatics are sensitive to aspirin and acidic NSAIDs.25 Ketorolac has been reported to cause acute bronchospasm in aspirin-intolerant patients.26 Systemic anaphylaxis associated with NSAID is very rare but is often fatal. There were 12 patients who experienced severe allergic reactions in our study. None of these patients died and no difference was found between ketorolac and the comparator NSAIDs.
The overall risk of serious adverse effects with ketorolac and the comparator NSAIDs was very low in this study. We found no evidence that the risk of serious adverse effects with ketorolac was any different to that with ketoprofen or diclofenac. It should be noted that these drugs were administered according to their approved dose and duration of treatment. We conclude that parenteral and oral ketorolac, ketoprofen, and diclofenac are equally safe for the treatment of pain after major surgery when administered according to the approved product label.
![]() |
Acknowledgements |
---|
![]() |
Appendix 1 |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Increased surgical site bleeding
An overt or covert loss of blood from the operative site that requires medical intervention and/or is excessive and/or is unexpected. For internal bleeding, evidence of shock, hypotension (<100 mm Hg systolic, <60 mm Hg diastolic), decreased haemoglobin and haematocrit, and/or increased blood urea and creatinine. For external bleeding, evidence of excessive bleeding into surgical drains and/or wound dressings, shock, hypotension, hypovolaemia, decreased haemoglobin and haematocrit, and/or increased blood urea and creatinine.
Gastrointestinal bleeding
An overt loss of blood from the gastrointestinal tract (haematemesis and/or melaena and/or blood per rectum) that requires medical intervention including endoscopy, transfusion of blood and/or i.v. fluids, or surgical reoperation to control bleeding, with evidence of shock, hypotension, decreased haemoglobin and haematocrit, and/or increased blood urea and creatinine.
Acute renal failure
Renal impairment that results in 100% increase in serum creatinine and/or oliguria, and/or dialysis, with evidence of increased blood urea and increased potassium, i.v. pyelogram, renal biopsy, x-rays, and/or ultrasound.
Severe allergic reaction
An allergic reaction that requires inhaled and/or i.v. medication, and/or intubation, and/or ventilation, with evidence of bronchspasm, dyspnoea, apnoea, angioedema, laryngeal oedema, shock, hypotension, anaphylaxis, skin rash, inotropic drugs, adrenergic drugs, antihistamines, and/or i.v. corticosteroids.
![]() |
Appendix 2 |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Souter AJ. Controversies in the perioperative use of nonsteroidal anti-inflammatory drugs. Anesth Analg 1994; 79: 117890[ISI][Medline]
3 Gales BJ, Gales MA. Death associated with inappropriate ketorolac dosing. Ann Pharmacother 1995; 29: 1299[ISI][Medline]
4 Fleming BM, Coombs DW. Bleeding diathesis after perioperative ketorolac. Anesth Analg 1991; 73: 235
5 Quan DJ, Kayser SR. Ketorolac-induced acute renal failure following a single dose. J Toxicol Clin Toxicol 1994; 32: 3059[ISI][Medline]
6 Summers K, Wilson R. The management of drug safety in the new Europe: ketorolac, a case study. Pharmaceutical Med 1994; 8: 13543
7 Strom BL, Berlin JA, Kinman JL, et al. Parenteral ketorolac and the risk of gastrointestinal and operative site bleeding. A postmarketing surveillance study. JAMA 1996; 275: 37682[Abstract]
8 Forrest J. Nonsteroidal anti-inflammatory drugs (NSAIDs). Anesth Clinics N Am: Ann Anesth Pharmacol 1997; 1: 179204
9 Langman MJ. Risks of anti-inflammatory drug-associated damage. Inflamm Res 1999; 48: 2368[ISI][Medline]
10 Kehlet H, Dahl JB. Are perioperative nonsteroidal anti-inflammatory drugs ulcerogenic in the short-term. Drugs 1992; 44 (Suppl 5): S3841
11 Willett LR, Carson JL, Strom BL. Epidemiology of gastrointestinal damage associated with nonsteroidal anti-inflammatory drugs. Drug Safety 1994; 10: 17081[ISI][Medline]
12 Garcia Rodriguez LA, Cattaruzzi C, Troncon MG, Agostinis L. Risk of hospitalization for upper gastrointestinal tract bleeding associated with ketorolac, other nonsteroidal anti-inflammatory drugs, calcium antagonists, and other antihypertensive drugs. Arch Intern Med 1998; 158: 339
13 Langman MJ. Epidemiology of nonsteroidal anti-inflammatory drug damage to stomach and duodenum. Ital J Gastroenterol Hepatol 1999; 33 (Suppl 1): S2S5
14 Bricker SRW, Savage ME, Hanning CD. Perioperative blood loss and non-steroidal anti-inflammatory drugs: an investigation using diclofenac in patients undergoing transurethral resection of the prostate. Eur J Anaesthesiol 1987; 4: 42934[ISI][Medline]
15 Wierod FS, Frandsen NJ, Jacobsen JD, Hartvigsen A, Olsen PR. Risk of haemorrhage from transurethral prostatectomy in acetylsalicylic acid and NSAID-treated patients. Scand J Urol Nephrol 1998; 321: 202
16 Conrad KA, Fagan TC, Mackie MJ, Mayshar PV. Effects of ketorolac tromethamine on hemostasis in volunteers. Clin Pharmacol Ther 1988; 43: 5426[ISI][Medline]
17 Antiplatelet Trialist Collaboration: Collaborative overview of randomised trials of antiplatelet therapy-1: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994; 308: 81106
18 Mismetti P, Laporte S, Darmon J-Y, Buchmüller A, Decousus H. Meta-analysis of low molecular weight heparin in the prevention of venous thromboembolism in general surgery. Br J Surg 2001; 88: 913930[ISI][Medline]
19 Greer IA, Gibson JL, Young A, Johnstone J, Walker ID. Effect of ketorolac and low-molecular weight heparin individually and in combination on haemostasis. Blood Coagul Fibrinolysis 1999; 10: 36773[ISI][Medline]
20 Sharma S, Chang DW, Koutz C, Evans GR, Robb GL, Langstein HN, Kroll SS. Incidence of hematoma associated with ketorolac after TRAM flap breast reconstruction. Plast Reconstr Surg 2001; 107: 3525[ISI][Medline]
21 Harris K. The role of prostaglandins in the control of renal function. Br J Anaesth 1992; 69: 2335[ISI][Medline]
22 Pearce CJ, Gonzalez FM, Wallin JD. Renal failure and hyperkalemia associated with ketorolac tromethamine. Arch Intern Med 1993; 153: 10002[Abstract]
23 Feldman HI, Kinman JL, Berlin JA, et al. Parenteral ketorolac: the risk for acute renal failure. Ann Intern Med 1997; 126: 1939
24 Lee A, Cooper MG, Craig JC, Knight JF, Keneally JP. The effects of nonsteroidal anti-inflammatory drugs (NSAIDs) on postoperative renal function: a meta-analysis. Anaesth Intens Care 1999; 27: 57480[ISI][Medline]
25 Power I. Aspirin-induced asthma. Br J Anaesth 1993; 71: 61921[ISI][Medline]
26 Chen AH, Bennett CR. Ketorolac-induced bronchospasm in an aspirin-intolerant patient. Anesth Prog 1994; 41: 1027