Secondary Logo

Journal Logo

Ketorolac is Not Nephrotoxic in Connection with Sevoflurane Anesthesia in Patients Undergoing Breast Surgery

Laisalmi, Merja MD; Eriksson, Heidi MD, PhD; Koivusalo, Anna-Maria MD, PhD; Pere, Pertti MD, PhD; Rosenberg, Per MD, PhD; Lindgren, Leena MD PhD

doi: 10.1097/00000539-200104000-00048
General Articles: Research Report

Ketorolac, which may cause renal vasoconstriction by cyclooxygenase inhibition, is often administered to patients anesthetized with sevoflurane that is metabolized to inorganic fluoride (F-), another potential nephrotoxin. We assessed this possible interaction using urine N-acetyl-β-D-glucosaminidase indexed to urinary creatinine (U-NAG/crea) as a marker of proximal tubular, β2-microglobulin as a tubular, urine oxygen tension (PuO2) as a medullary, and erythropoietin as a marker of tubulointerstitial damage. Thirty women (ASA physical status I-II) undergoing breast surgery were included in our double-blinded study. They were allocated into two groups receiving either ketorolac 30 mg IM (Group K) or saline (Group C) at the time of premedication, at the end of, and 6 h after anesthesia maintained with sevoflurane. Urine output, U-NAG/crea, PuO2, serum creatinine, urea, and F- were assessed. Blood loss was larger in Group K (465 ± 286 mL vs 240 ± 149 mL, mean ± sd, P < 0.05). The MAC-doses of sevoflurane were similar. U-NAG/crea increased during the first 2 h of anesthesia and serum F- peaked 2 h after the anesthesia without differences between the groups. There were no statistically significant changes in PuO2, erythropoietin, β2-microglobulin, serum creatinine, urea, or urine output during anesthesia or the recovery period in either group. Our results indicate that the kidneys are not affected by ketorolac administered in connection with sevoflurane anesthesia.

Department of Anaesthesiology and Intensive Care Medicine, Helsinki University Hospital, Helsinki, Finland

Supported, in part, by Helsinki University Central Hospital EVO Grant TKI4009.

December 7, 2000.

Address correspondence and reprint requests to Merja Laisalmi, MD, Department of Anaesthesiology and Intensive Care Medicine, Helsinki University Hospital, Surgical Hospital, PO BOX 263, 00029 HUS, Helsinki, Finland. Address e-mail to

IMPLICATIONS: The different kinetics of N-acetyl-β-D-glucosaminidase indexed to urinary creatinine and serum inorganic fluoride during and after sevoflurane anesthesia suggest that the observed mild renal tubular function deterioration is not caused by inorganic fluoride. Administration of ketorolac IM is therefore considered safe in adequately hydrated healthy adult patients given sevoflurane anesthesia.

Inorganic fluoride (F-) released in the metabolism of sevoflurane may reach serum nephrotoxic concentrations after methoxyflurane anesthesia (1). Despite this, clinical renal dysfunction after sevoflurane anesthesia is a very rare complication (2). However temporary tubular dysfunction, as indicated by the sensitive urinary markers, N-acetyl-β-D-glucosaminidase (NAG) and β2-microglobulin (β2-M), has been reported after sevoflurane anesthesia (3).

Nonsteroidal antiinflammatory drugs (NSAIDs) are commonly given intraoperatively to control postoperative pain (4). By blocking the cyclooxygenase enzymes, the NSAIDs also block the synthesis of prostaglandin E2 and I2, which are important renal vasodilators (5). Clinically, this renal vasoconstrictive effect of NSAIDs is reflected in decreased urine output in the postoperative period (6).

We assume that the combination of two potential renal toxins could be harmful to the kidneys. Sensitive markers of both renal tubular damage and ischemic renal insult were used in the present study to detect even subclinical renal toxicity when ketorolac was given in combination with sevoflurane anesthesia.

Back to Top | Article Outline


The double-blinded, placebo-controlled study protocol was approved by the ethics committee of the hospital. Thirty ASA physical status I-II women scheduled to undergo elective breast surgery gave written informed consent. Patients with abnormal renal or hepatic function were excluded. Serum creatinine, glutamyl transferase, alkaline phosphatase, sodium, and potassium were used to test preoperative renal or hepatic function. All patients had inpatient surgery. They fasted from midnight prior to the day of the surgery and no additional fluid administration was given during that time. The fasting period was less than 12 h in all patients. Diazepam (0.2 mg kg−1) was given orally for premedication. The patients were divided into two groups. The ketorolac group (Group K) received 30 mg ketorolac IM (IM) with the premedication, at the end of, and 6 h after anesthesia. The control group (Group C) received three IM injections of saline.

The patients were anesthetized with propofol 2 mg · kg−1, fentanyl 2 μg · kg−1, and glycopyrrolate 0.2 mg IV. Tracheal intubation was facilitated with cisatracurium 0.15 mg · kg−1. Anesthesia was maintained with sevoflurane in oxygen-air mixture (Fio2 0.4). Ventilation was controlled with fresh gas flow of 4–6 L/min using a semiclosed system (Engström ECS 20 ; Engström, Stockholm, Sweden) to maintain end-tidal (ET) carbon dioxide tension (ETco2) at 4.5–5.0 kPa. The end-tidal concentration of sevoflurane was analyzed with a Capnomac Ultima capnograph (Datex, Helsinki, Finland), which was calibrated immediately before anesthesia. The concentration of sevoflurane was adjusted to maintain mean arterial pressure (MAP) within 20% of baseline and above a minimum of 60 mm Hg. The MAC hours of sevoflurane were calculated from the ET concentration of the anesthetic and duration of the exposure. If the heart rate increased more than 30% of the baseline value or over 90 bpm and did not respond to an increased concentration of sevoflurane, a bolus of alfentanil 0.5 mg IV was given.

The surgical blood loss was measured by weighing the cotton swabs used during surgery. The amount of blood suctioned from the surgical area was measured. A 70-cm catheter was inserted via the basilic vein to the superior caval vein to monitor central venous pressure (CVP) and to obtain blood samples. Correct placement of the catheter was verified recording intraatrial electrocardiogram using AlphaCard®(7) (B. Braun, Melsungen, Germany). Ringer’s acetated solution (Ringer-Acetat®, Pharmacia 38 Upjohn, Stockholm, Sweden) and 6% hydroxyethyl starch solution (Plasmafusin 6%®, Kabi Pharmacia, Sweden) were used to maintain CVP at 6–12 mm Hg. A urinary bladder catheter was inserted to measure urine output and to collect urine samples. If the urine output was <0.5 mL/kg/h and CVP was less than 6 mm Hg, an additional bolus of 300–500 mL of Ringer’s acetated solution was given.

Samples for the analyses of serum and urine fluoride, β2-M, and urine NAG/crea (units of urinary NAG activity per gram of urinary creatinine) were collected preoperatively, after 2 h of anesthesia, 2 and 12 h after the end of anesthesia, as well as on the first and on the second postoperative days (PODs). Samples for the assays of serum and urine sodium, potassium, osmolality, and serum erythropoietin (EPO), serum urea, and creatinine concentrations were taken preoperatively, 12 h after the end of anesthesia, and on the first and second POD. Hemoglobin and hematocrit were measured preoperatively, on the first POD and additionally when needed. Urine oxygen tension (PuO2) was determined every 20 min in the operating room and in the postanesthesia care unit (PACU).

The concentrations of F- were determined by a method modified from that of Fry and Taves (8). A fluoride selective combination electrode Orion model 96–09(Orion Research Incorporated, Boston, MA) was used for the measurement on Parafilm “M” (American National Can, Greenwich, CT) placed on 16-mm cell culture wells. Before measurement, 200 μL of acetate buffer (acetate-NaOH 1 mol · L−1, pH 5.2, NaCl 1 mol · L−1), and 10 μL of sodium fluoride 20 μmol · L−1 were added to 190 μL of serum.

The activity of U-NAG was determined by using 3-cresonsulphonphtalein-N-acetyl-β-glucosamine as a substrate (Boehringer Mannheim Biochemica, Mannheim, Germany). The method was adapted to a Kone Specific® random access analyzer (Kone Instruments, Helsinki, Finland). Kinetic follow-up of the reaction was used instead of determination of end point absorbance. Urine NAG and creatinine values were measured from spot samples. The activity of NAG was normalized to creatinine concentration and expressed as U-NAG/crea (9).

Urine oxygen tension as a marker of renal medullary homeostasis was measured with a blood gas analyzer (Synthesis 35®; Instrumentation Laboratory, Laboratories Scandinavia, Helsinki, Finland) from the samples taken from fresh urine in the urinary catheter. EPO, a marker of tubulointerstitial function was analyzed using an in-house radioimmunologic method with reagents from Medix Diacor (Espoo, Finland) and an international reference standard (second international reference preparation 67/343), and serum and urine β2-M samples were analyzed using time-resolved fluoroimmunoassay with dissociation enhanced lanthanide fluoroimmunoassay (DELFIA) β2- microkit® (Wallac, Turku, Finland). The analyses of serum and urine sodium, potassium and creatinine concentrations as well as serum and urine osmolality and serum urea concentration were performed in the clinical laboratory of the hospital.

Postoperatively the patients were observed for two hours in the PACU. Heart rate, MAP and CVP were measured during the anesthesia and the PACU stay. Postoperative pain was scored using a visual analog scale from 1–10, and if the patient scored ≥ 4 she was given morphine 0.05 mg · kg−1 IV. Before discharge to the ward the patients were given a patient-controlled analgesia device programmed to deliver a dose of morphine 2 mg IV when required, with a lock-out time of 8 min and no more than four doses per hour. Postoperative nausea was treated with droperidol and ondansetron as needed.

Data within a group were analyzed using one-way analysis of variance. For differences between the two groups, two-way analysis of variance for repeated measures or Student’s unpaired t-test were used. Fisher’s protected least significance difference test was used to test significant changes. Calculations were performed using Stat View 5.0.1(SAS Institute, San Francisco, CA). Data are expressed as mean ± sd (tables) or sem (figures). A probability value of <0.05 was considered statistically significant.

Back to Top | Article Outline


The characteristics of the patients were comparable except that the surgical blood loss was larger in the Group K than in Group C (P < 0.05) (Table 1). The number of patients undergoing radical mastectomy and axillary dissection was four in Group K and five in Group C. The rest of the patients underwent minor breast surgery. The bleeding in patients with radical mastectomy and axillary dissection was 800 ± 355 mL in Group K and 330 ± 186 mL in Group C (P < 0.05). The bleeding in patients undergoing minor breast surgery was 340 ± 124 mL in Group K and 200 ± 103 mL in Group C (P < 0.05).

Table 1

Table 1

Serum F- levels increased moderately in all groups and returned to the baseline during the first 24 postoperative hours. The peak serum fluoride levels were measured 2 h after the end of the anesthesia. Urine fluoride concentration was similar in both groups (Fig. 1). U-NAG/crea increased earlier (i.e., 2 h after the induction of the anesthesia in both groups) and returned to the baseline during the first 24 h after the anesthesia (Fig. 2). PuO2 remained unchanged in both groups during the operation and PACU stay (Table 2).

Figure 1

Figure 1

Figure 2

Figure 2

Table 2

Table 2

Serum EPO concentrations increased in both groups until 24 h postoperatively and remained at that level during the follow-up period (P < 0.01) (Table 2). Blood hemoglobin concentrations and hematocrit decreased significantly in both groups after the operation (P < 0.05) (Table 3).

Table 3

Table 3

Serum β2-M levels remained within the reference limits of our laboratory (0.6–3.0 mg/L) (Table 2). Urine β2-M remained lower than 200 mg/L, which is the threshold for the measurement. Occasional individual values greater than 200 mg/L were measured (data not shown).

Urine sodium concentration decreased in Group K on the first POD (P < 0.05 compared with the preanesthetic value) and returned to the preanesthetic level on the second POD. Serum sodium concentration remained within the normal range during the study period. There was a significant decrease in serum potassium in Group C (P < 0.05) although the concentrations remained within the reference limits. Serum osmolality decreased in both groups after 12 h of anesthesia and returned to the preanesthetic levels, urine osmolality remaining reduced during the first and second POD. Serum creatinine and urea decreased in both groups during the first 24 h after anesthesia (Table 3).

There were no significant differences in CVP or urine output during the perioperative period (Table 4).

Table 4

Table 4

Back to Top | Article Outline


In our study, renal function was assessed using sensitive markers that monitor the function of different entities of the kidney. U-NAG/crea for proximal tubule, PuO2 is a marker of medullary homeostasis, and EPO that of the tubulointerstitium. The traditional function markers such as serum creatinine and urea were also measured.

The renal toxic threshold for serum F- concentration is considered to be 50 μmol/L based on the experiences of methoxyflurane anesthesia (1). Methoxyflurane is metabolized to F- in the kidneys to a significantly larger extent than sevoflurane (2). This may be one reason that, in studies with sevoflurane, serum F- concentrations exceeding 100 μmol/L have been measured with no clinical renal deterioration. Compound A is another fluorine containing product of reaction between sevoflurane and desiccated carbon dioxide (CO2) absorbent (10). Compound A is toxic in animal tests (11) but there have been no clinical signs of renal toxicity in humans (12). In our study, moderately high fresh gas flows without a CO2 absorber were used, thus eliminating the role of Compound A.

Nonselective cyclooxygenase enzyme-blocking NSAIDs have caused vasoconstriction in the renal vasculature (5), especially renal arterioles. The inhibition of prostaglandin synthesis may compromise renal blood flow and glomerular filtration rate in connection with the effects of surgical stress. These harmful effects have been seen in elderly and hypovolemic patients in connection with hypotensive periods during anesthesia (13). Adequate volume loading protects renal function in connection with the use of radiographic contrast media (14). In our study, special attention was paid to prevention of dehydration and hypotension, the well-known risk factors of NSAID-induced renal damage.

Higuchi et al. (3) showed a significant increase in U-NAG/crea ratio in patients with serum F- concentrations exceeding 50 μmol/L after sevoflurane anesthesia. This change occurred on the second POD. They did not control the intravascular fluid administration of the patients who also received potentially nephrotoxic antibiotics. Intraoperatively, a tourniquet on the thigh or arm was used in the majority of the patients. The unavoidable ischemic and compression tissue injury may also have harmful effects on renal function (15). Thus several factors possibly affecting renal function were present.

We noted an increase in the U-NAG/crea in both groups when the patients had been two hours under anesthesia. At that time the urine NAG/crea was increased. All of the patients were exposed to surgery and 15 of them also to the first dose of ketorolac. The peak serum concentrations of F-, however, were detected two hours after the end of anesthesia. The change in U-NAG/crea returned to the baseline during the first POD. Thus, the high F- concentrations or the administration of ketorolac cannot be the reason for NAG release in the urine. A recent study shows that NAG excreted in urine is a marker of deteriorated tubular function and a marker of increased lysosomal activity in the tubular cells, and not necessarily an indicator of cell death (16). This small increase in U-NAG/crea might be considered a sign of disturbed functional integrity of the tubular cells. The stress attributable to the anesthesia and surgery may explain the increased U-NAG/crea. This speculation warrants further studies.

In our patients PuO2 remained at the preoperative level during the surgery and PACU stay, indicating adequate renal medullary homeostasis in both groups. Hypoxia and hemorrhage stimulate EPO synthesis in humans. In an animal study, ketorolac had no effect on EPO production and release in response to reduced hematocrit (17). An increase in EPO levels paralleled the slight reduction of hemoglobin concentration in our study and the administration of ketorolac had no effect on EPO production. This also indicates that there was no tubulointerstitial renal damage.

Some investigations have shown increases in urinary β2-M after anesthesia (18,19). In our study changes were seen neither in serum nor urine β2-M, indicating intact tubular reabsorption and glomerular filtration.

Serum osmolality decreased minimally 12 hours postoperatively but returned to the preoperative level irrespective of urine osmolality, which remained low for the first two days. The hypoosmolar diuresis was the logic consequence of our active hydration policy. Serum osmolality remained stable because of the homeostatic control of serum osmolality by the central nervous system and kidneys.

We did not note any differences in PuO2, serum or urine β2-M, U-NAG/crea or EPO between the groups. These variables are very sensitive to react to any adverse renal changes. Their use in clinical practice and impact on outcome are under debate. A recent report of renal effects of sevoflurane (20) showed that 3463 patients who underwent sevoflurane anesthesia had no changes in serum creatinine or blood urea nitrogen. This was also the case in our study. From a practical point of view, creatinine and urea can be considered appropriate markers of renal function and clinical outcome when assessing the renal effect of modern inhaled anesthetics.

Intraoperative blood loss was larger in the ketorolac group, which is in agreement with an earlier study of different kinds of surgery (21). However, bleeding after mastectomy in patients given ketorolac IV intraoperatively has not been reported to be more than in a placebo group (4). The lack of an effect of ketorolac on blood loss in the study by Bosek and Cox (4) may be a result of the fact that ketorolac was administered near the end of the surgery. The bolus dose of ketorolac given to the patients was the same as our study.

Day-case surgery has become very popular. The patients are older and may have deteriorated renal function influencing the outcome. With a safe combination of a fast-acting anesthetic and an effective nonopioid analgesic, the number of patients needing hospitalization after day surgery may be reduced. In light of our data, the combination of sevoflurane and ketorolac seems to be safe also in day-case surgery.

In conclusion, in otherwise healthy patients undergoing breast surgery under sevoflurane anesthesia, we noted only a slight increase in U-NAG/crea before the peak serum F- concentrations. The administration of ketorolac 90 mg IM perioperatively within approximately 10 hours did not influence serum creatinine and urea values, nor any of the sensitive renal cellular function markers. In well hydrated patients the combination of ketorolac and moderate doses of sevoflurane (3–4 MAC hours) appears to be safe to the kidneys.

Back to Top | Article Outline


1. Cousins MJ, Mazze RI. Methoxyflurane nephrotoxicity: a study of dose response in man. JAMA 1973; 225: 1611–6.
2. Kharasch ED, Hankins DC, Thummel KE. Human kidney methoxyflurane and sevoflurane metabolism: intrarenal fluoride production as a possible mechanism of methoxyflurane nephrotoxicity. Anesthesiology 1995; 82: 689–99.
3. Higuchi H, Sumikura H, Sumita S, et al. Renal function in patients with high serum fluoride concentrations after prolonged sevoflurane anesthesia. Anesthesiology 1995; 83: 449–58.
4. Bosek V, Cox CE. Comparison of analgesic effect of locally and systemically administered ketorolac in mastectomy patients. Ann Surg Oncol 1996; 3: 62–6.
5. Dunn MJ, Scharschmidt L, Zambraski E. Mechanisms of the nephrotoxicity of non-steroidal anti-inflammatory drugs. Arch Toxicol Suppl 1984; 7: 328–37.
6. Perttunen K, Kalso E, Heinonen J, Salo J. IV diclofenac in post-thoracotomy pain. Br J Anaesth 1992; 68: 474–80.
7. Salmela L, Aromaa U. Verification of the position of a central venous catheter by intra- atrial ECG: when does this method fail? Acta Anaesthesiol Scand 1993; 37: 26–8.
8. Fry BW, Taves DR. Serum fluoride analysis with the fluoride electrode. J Lab Clin Med 1970; 75: 1020–5.
9. Wellwood JM, Ellis BG, Price RG, et al. Urinary N-acetyl-β-D-glucosaminidase activities in patients with renal disease. BMJ 1975; 3: 408–11.
10. Steffey EP, Laster MJ, Ionescu P, et al. Dehydration of Baralyme increases compound A resulting from sevoflurane degradation in a standard anesthetic circuit used to anesthetize swine. Anesth Analg 1997; 85: 1382–6.
11. Kharasch ED, Thorning D, Garton K, et al. Role of renal cysteine conjugate beta-lyase in the mechanism of compound A nephrotoxicity in rats. Anesthesiology 1997; 86: 160–71.
12. Kharasch ED, Frink EJ Jr, Zager R, et al. Assessment of low-flow sevoflurane and isoflurane effects on renal function using sensitive markers of tubular toxicity. Anesthesiology 1997; 86: 1238–53.
13. Eknoyan G. Current status of chronic analgesic and nonsteroidal anti-inflammatory drug nephropathy. Curr Opin Nephrol Hypertens 1994; 3: 182–8.
14. Solomon R, Werner C, Mann D, et al. Effects of saline, mannitol, and furosemide to prevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med 1994; 331: 1416–20.
15. Orsi R, Brunelli G, Zorzi F. Lesions of the kidney in tourniquet shock: ultrastructural study. Microsurgery 1984; 5: 191–6.
16. Bosomworth MP, Aparicio SR, Hay AW. Urine N-acetyl-β-D-glucosaminidase - a marker of tubular damage? Nephrol Dial Transplant 1999; 14: 620–6.
17. Hoff KK, Zawada ET, Alavi FK, et al. Effects of ketorolac tromethamine on erythropoietin levels in Sprague Dawley rats. Int J Artif Organs 1994; 17: 629–34.
18. Nishiyama T, Hanaoka K. Inorganic fluoride kinetics and renal and hepatic function after repeated sevoflurane anesthesia. Anesth Analg 1998; 87: 468–73.
19. Higuchi H, Sumita S, Wada H, et al. Effects of sevoflurane and isoflurane on renal function and on possible markers of nephrotoxicity. Anesthesiology 1998; 89: 307–22.
20. Mazze RI, Callan CM, Galvez ST, et al. The effects of sevoflurane on serum creatinine and blood urea nitrogen concentrations: a retrospective, twenty-two-center, comparative evaluation of renal function in adult surgical patients. Anesth Analg 2000; 90: 683–8.
21. Strom BL, Berlin JA, Kinman JL, et al. Parenteral ketorolac and risk of gastrointestinal and operative site bleeding: a postmarketing surveillance study. JAMA 1996; 275: 376–82.
© 2001 International Anesthesia Research Society