Secondary Logo

Journal Logo

Original Article

COX 2 selectivity of non-steroidal anti-inflammatory drugs and perioperative blood loss in hip surgery. A randomized comparison of indomethacin and meloxicam

Weber, E. W. G.*; Slappendel, R.; Durieux, M. E.*; Dirksen, R.; van der Heide, H.; Spruit, M.

Author Information
European Journal of Anaesthesiology: December 2003 - Volume 20 - Issue 12 - p 963-966


Non-steroidal anti-inflammatory drugs (NSAIDs) are used in the perioperative period for analgesia and reduction of oedema in the surgical field. However, they exhibit several side-effects: reduction of renal blood flow, an increased incidence of gastric complaints and increased blood loss during surgery. These adverse effects result from inhibition of the physiological formation by cyclo-oxygenase (COX) 1 of several prostanoids: PGE2 and PGI2 with a cytoprotective function, and thromboxane A2, which is responsible for platelet aggregation. In contrast, newer COX 2-selective NSAIDs, such as meloxicam, are considered to influence only the inflammatory response (oedema, pain and fever). Our hypothesis was that this selectivity of action can be extrapolated to clinical practice, and that administration of a COX 2-selective NSAID would result in decreased perioperative blood loss compared with administration of a non-selective compound. To test this hypothesis, we performed a randomized study in patients undergoing total hip replacement, and compared perioperative blood loss after use of one of two NSAIDs: indomethacin and meloxicam.


The study was approved by the Ethical Committee of our Hospital. Patients (ASA I-III) scheduled for total hip surgery using intrathecal anaesthesia were potentially eligible for the study. Patients with a history of peptic ulcer disease, renal dysfunction or allergy to any NSAID were excluded. The remaining patients were randomized to one of two groups: those receiving meloxicam 15 mg orally, once a day, or those receiving indomethacin 50 mg orally, three times a day. The study was not masked. Patients were asked to stop use of any NSAID 2 weeks prior to surgery. In case of severe pain acetaminophen (paracetamol) was available if necessary. Patients who failed to stop the use of NSAIDs were excluded.

All patients were premedicated with midazolam 7.5 mg orally 1 h before administration of intrathecal anaesthesia; at the same time the first dose of NSAID was given. Intrathecal anaesthesia (27-G pencil point needle) was administered using bupivacaine 20 mg plus morphine 0.1 mg to a total volume of 4 mL. The anaesthesiologist administered midazolam (1 mg at intervals of not less than 5 min) until the patient indicated that adequate sedation was achieved. Fluid replacement was by protocol: glucose 500 mL 2.5%/NaCl 0.9% solution was administered after placement of an intravenous cannula before starting surgery. The same solution was infused during surgery at 250 mL h−1, and after surgery at 100 mL h−1. A colloid solution (Gelofusine®) was administered to match measured blood loss (see below), with a maximum of 2.5 L per 24 h. Non-invasive arterial pressure, heart rate (electrocardiogram), oxygen saturation (SPO2), and respiratory frequency were continuously monitored during anaesthesia and in the intensive care unit (ICU) during the first 24 h after surgery.

Blood loss

One day before surgery, patients received acenocoumarol 3 mg. Intraoperative blood loss was measured by operating nurses unaware of the NSAID given. Total blood loss was calculated by taking into account the volume in the suction containers, the weight of the surgical sponges, and the irrigation fluid used. In the postoperative phase all patients remained in an ICU for 24 h, which is our normal policy. The volume of blood in a high vacuum wound drainage system was collected and measured for 24 h after surgery.


In the postoperative period, all patients were treated with indomethacine 50 mg orally, three times a day, or meloxicam 15 mg orally, once a day. If pain was present morphine was administered intravenously by patient controlled analgesia (PCA) pump (BRAUN, Melsungen, Germany). The settings of the PCA pump were: basal rate 0.0 mg h−1, bolus dose 1.0 mg, bolus interval 5 min, maximum 30 mg per 4 h. Pain was evaluated using visual analogue scores (VAS, ranging from 0 to 10: 0 = no pain, 10 = most severe pain). For each individual patient we assessed the maximum VAS score in the 24 h period and cumulative VAS scores, quantified as area under the curve (AUC) of VAS scores during the 2 h period.

Statistical analysis

To detect a difference of 100 mL of blood loss during surgery (SD ± 250 mL) with an α error (two sided) of 0.05 and a β error of 0.10, it was necessary to include 84 patients per group.

Analysis of interval scored data was performed using the Student's t-test. Non-parametric techniques (Kruskal-Wallis test) were used when necessary. Proportions were analysed with χ2 statistics and Fisher's exact test. The α level for all analyses was set at P = 0.05. Data are reported as mean ± SD.


In total 200 consecutive patients consented to participate during the preoperative outpatient visit. However, at the time of admission 32 patients had failed to stop use of regular NSAIDs, mainly because of pain. Of the remaining 168 patients, 82 were randomized to indomethacin and 86 were randomized to meloxicam. Patient characteristics data of these remaining patients are provided in Table 1. The two groups were similar with regard to age, height, weight, gender, mean arterial pressure and heart rate.

Table 1
Table 1:
Patient characteristics and intraoperative factors.

Blood loss

Blood loss was significantly greater in patients pretreated with indomethacin than in those receiving meloxicam (Table 2). The volume of blood loss during operation was 19% greater in the indomethacin group than in the meloxicam group (P < 0.01). Likewise, blood loss in the first 24 h after surgery was 15% greater in the indomethacin group than in the meloxicam group (P < 0.05). As a result, total blood loss (i.e. intraoperative blood loss plus blood loss during the first 24 h after surgery) was 17% greater in the indomethacin group than in the meloxicam group (P < 0.05).

Table 2
Table 2:
Intra- and postoperative blood loss.


VAS scores were <3 in the postoperative period in all patients in both groups (Fig. 1). Cumulative VAS scores were 6.1 for the indomethacin group and 5.3 for the meloxicam group (P > 0.05). Use of systemic morphine by PCA infusion pump was similar: 18 ± 9 mg in the meloxicam group and 18 ± 7 mg in the indomethacin group.

Figure 1
Figure 1:
VAS pain scores 24 h after surgery. ♦: Meloxicam; ▪: Indomethacin.


The major finding of our study is that pretreatment with the COX 2-selective NSAID meloxicam results in less perioperative blood loss than use of the non-selective compound indomethacin. This data from clinical practice is consistent with in vitro studies that demonstrated the COX 2-selective action of meloxicam. However, several limitations of the study should be kept in mind. We did not stratify patients by surgeon, as routine audit data from our hospital show very similar amounts of blood loss among the surgical staff. However, this is a potential confounder. In addition, we did not determine if the observed effects on blood loss had a significant impact on patient outcome or management (e.g. haemoglobin concentration or transfusion requirements).

NSAIDs are used widely in orthopaedic surgery, but concerns remain about their tendency to increase intraoperative blood loss [1-3]. This side-effect results from impairment of platelet aggregation by block of thromboxane formation, and could not be explained by an interaction between indomethacin and acenocoumarol [4,5]. A breakthrough in the ability to separate the beneficial anti-inflammatory, analgesic and antipyretic action of NSAIDs from these side-effects resulted from better insights in their mechanism of action. NSAIDs inhibit COX, the major enzyme in the biosynthesis of prostaglandins. In the early 1990s, it was recognized that the rate of prostaglandin synthesis could increase dramatically when the formation of a particular isoform of COX is induced by several cytokines or lipopolysaccharide. It is now known that COX exists in at least two isoforms, known as COX 1 and COX 2 [6,7]. COX 1 is found in the stomach, intestine, kidneys and platelets, and is essential for the synthesis of prostaglandins involved in important physiological processes, such as protection of the gastrointestinal mucosa, maintenance of renal function and circulatory homeostasis. This role of COX 1 has been referred to as the 'housekeeping' function. In contrast, the inducible isoform COX 2 produces large amounts of prostaglandins that mainly contribute to the pathophysiological process of inflammation. Thus, the therapeutic effects of NSAIDs are largely the result of inhibition of the enzyme COX 2, whereas adverse effects are primarily due to the inhibition of COX 1.

It has been suggested that NSAIDs that selectively inhibit COX 2 have fewer side-effects [6,8]. An example is meloxicam, a NSAID derived from enolic acid which has a favourable COX 2/COX 1 ratio [9,10]. This ratio translates into fewer effects on platelet aggregation in vitro. Indomethacin is a COX 2/COX 1 non-selective NSAID when tested in vitro, but shows a slight preference for COX 2 when tested ex vivo[11]. Studies in vivo show contradictory results on the effects of indomethacin on blood loss [12,13]. In this study we compared two different NSAIDs, which inhibit the two COX-isoforms to varying degrees.

More relevant to clinical practice is the question whether these data from in vitro studies and animal experiments translate in a decreased volume of blood loss. The present study indicates that such is indeed the case: blood loss was approximately 20% less when meloxicam instead of indomethacin was used. However, whether this is of clinical significance and influences patient outcome remains to be determined. In addition, the potential beneficial effects of these compounds on blood loss should be weighed against potential detrimental effects (such as a potentially increased risk for cardiovascular events) before routine use can be recommended.


1. Campbell WI, Kendrick R, Patterson C. Intravenous diclofenac sodium. Does its administration before operation suppress postoperative pain? Anaesthesia 1990; 45: 763-766.
2. Rorarius MG, Baer GA, Metsa-Ketela T, Miralles J, Palomaki E, Vapaatalo H. Effects of peri-operatively administered diclofenac and indomethacin on blood loss, bleeding time and plasma prostanoids in man. Eur J Anaesthesiol 1989; 6: 335-342.
3. Schafer AI. Effects of nonsteroidal antiinflammatory drugs on platelet function and systemic hemostasis. J Clin Pharmacol 1995; 35: 209-219.
4. Michot F, Ajdacic K, Glaus L. A double-blind clinical trial to determine if an interaction exists between diclofenac sodium and the oral anticoagulant acenocoumarol (nicoumalone). J Int Med Res 1975; 3: 153-157.
5. Pardo A, Garcia Losa M, Fernandez Pavon A, et al. A placebo-controlled study of interaction between nabumetone and acenocoumarol. Br J Clin Pharmacol 1999; 47: 441-444.
6. McKenna F. COX-2: separating myth from reality. Scand J Rheumatol Suppl 1999; 109: 19-29.
7. Vane JR, Botting RM. New insights into the mode of action of anti-inflammatory drugs. Inflamm Res 1995; 44: 1-10.
8. Meade EA, Smith WL, DeWitt DL. Differential inhibition of prostaglandin endoperoxide synthase (cyclooxygenase) isozymes by aspirin and other non-steroidal anti-inflammatory drugs. J Biol Chem 1993; 268: 6610-6614.
9. de Meijer A, Vollaard H, de Metz M, Verbruggen B, Thomas C, Novakova I. Meloxicam, 15 mg/day, spares platelet function in healthy volunteers. Clin Pharmacol Ther 1999; 66: 425-430.
10. Stichtenoth DO, Wagner B, Frolich JC. Effects of meloxicam and indomethacin on cyclooxygenase pathways in healthy volunteers. J Investig Med 1997; 45: 44-49.
11. Giuliano F, Warner TD. Ex vivo assay to determine the cyclooxygenase selectivity of non-steroidal anti-inflammatory drugs. Br J Pharmacol 1999; 126: 1824-1830.
12. Perttunen K, Kalso E, Heinonen J, Salo J. IV diclofenac in post-thoracotomy pain. Br J Anaesth 1992; 68: 474-480.
13. Power I, Chambers WA, Greer IA, Ramage D, Simon E. Platelet function after intramuscular diclofenac. Anaesthesia 1990; 45: 916-919.

ANALGESICS, NON-NARCOTIC, anti-inflammatory agents, non-steroidal; ARTHROPLASTY, REPLACEMENT, HIP; HAEMORRHAGE, blood loss, surgical

© 2003 European Academy of Anaesthesiology