Propofol versus isoflurane for endoscopic sinus surgery

Propofol Versus Isoflurane Sinus Surgery

for Endoscopic

Janet D. Pavlin, MD, Peter S. Colley, MD, Ernest A. Weymuller Jr, MD, Gail Van Norman, MD, Holly C. Gunn, MD, and Meagan E. Koerschgen, BS Purpose: A previous retrospective study reported that propofol anesthesia decreased bleeding during endoscopic sinus surgery compared with isoflurane. We performed a prospective study to compare the effects of propofol versus isoflurane on measured blood loss and the surgeon’s subjective assessment of operating conditions during endoscopic sinus surgery. Patients and Methods: After receiving institutional review board approval and written informed consent, 56 patients undergoing endoscopic sinus surgery were randomly assigned to receive propofol (n = 30) or isoflurane (n = 26) supplemented with nitrous oxide-oxygen and alfentanil. Blood loss was calculated from the hemoglobin concentration in suction canisters. One surgeon, who was blinded to the anesthetic agent, performed every procedure and assessed bleeding as follows: 1, no bleeding; 2, modest bleeding; 3, bleeding interfering with operating conditions and cause for an agent switch; and 4, intolerable bleeding requiring a change in surgical plan. Results were compared in the two anesthetic groups using chi-squared test, unpaired t-test, Mann-Whitney Utest, and a permutation test. A Pof .05 was considered significant. Results: Mean bleeding scores were less over time (P = .02) with propofol anesthesia, particularly in surgery in the ethmoid and sphenoid sinuses (P = .03), and the proportion of patients with a mean score >2 was less in the propofol group (30% ~54%; P = .033). Time until discharge to home or to a limited stay in a hospital bed was also less in the propofol group (183 ~243 minutes; P = .019). Conclusion: In our study, surgical blood loss was the same for both anesthetic agents overall, but propofol appeared to offer an advantage in terms of subjective improvement in operating conditions, particularly in the ethmoid and sphenoid sinuses. (Am J Otolaryngol 1999:20:96-l 01. Copyright 0 1999 by W.B. Saunders Company)

Endoscopic surgery has revolutionized the surgical treatment of chronic sinus disease, permitting outpatient sinus surgery with minimal morbidity. Although sometimes performed using local anesthesia, the majority of patients prefer to have a general anesthetic for endoscopic sinus surgery. A critical factor in the management of general anesthesia is to provide a relatively bloodless field to optimize visibility for the surgeon. This is particularly important in the ethmoid and sphenoid sinuses, in which even minimal bleeding may seriously impair the ability to complete the intended surgery. From the Departments of Anesthesiology and Otolaryngology, University of Washington School of Medicine, Seattle, WA. Address reorint reauests to Janet D. Pavlin, MD, Box 356540, Department of Anesthesiology, University of Washinaton. Seattle. WA 98105. Copy;ighto 1999 by W.B. Saunders Company 0196-0709/99/2002-0004$10.00/0 96

American

Journal

of Otolaryngology,

Vol20,

Techniques commonly used to minimize bleeding include local injection of vasoconstrictors, head elevation, and controlled hypotension. Recently, Blackwell et all suggested that propofol anesthesia decreased the amount of bleeding during endoscopic sinus surgery compared with isoflurane. In that retrospective study of 25 patients, blood loss was estimated by an unblinded surgeon and was reported to be less in patients anesthetized with propofol compared with isoflurane (101 v 251 mL; P < .Ol). However, patients were not randomly assigned to treatment groups, and patients in the isoflurane group were, on average, 13 years older than those in the propofol group. The present study was designed to compare a surgeon’s subjective assessment of operating conditions and the measured rate of blood loss prospectively in patients randomly assigned to receive propofol or isoflurane anesthesia. No 2 (March-April),

1999:

pp 96-l 01

PROPOFOL

AND

SINUS

SURGERY

97

PATIENTS AND METHODS

lidocaine with l:lOO,OOO epinephrine into the submucosa at the anterior attachment of the middle turbinate and in the region of the sphenopalatine foramen on the side to be operated on first. When appropriate, injections were performed on the opposite side immediately before surgery on the second side. At I5-minute intervals, the surgeon provided numeric assessments of the operative conditions, primarily in relation to the amount of bleeding and its effect on visibility, and identified the sinus being treated. The scoring system used is shown in Table 1. Scores that were midway between the defined scores (ie, 1.5, 2.5) were accepted. If a score of >3 was attained, the anesthetic agent was switched to the alternate agent being studied. MAP, heart rate, oxyhemoglobin saturation, end-tidal gas concentrations (isoflurane, NzO, carbon dioxide), and concentrations of isoflurane being delivered were recorded at 15-minute intervals. The surgery and infiltration of local anesthetic was performed by a single surgeon who was blinded to the anesthetic being administered. Patients and recovery room nurses were unaware of which anesthetic had been administered. Near the end of surgery, the muscle relaxant was reversed by intravenous neostigmine, 3.0 mg, and glycopyrrolate, 0.6 mg. Alfentanil was stopped 15 minutes and propofol or isoflurane was stopped 5 minutes before the end of surgery; NzO was stopped at the end of surgery. The total time to discharge readiness and actual time of discharge were recorded.

Approval was obtained from the Human Subjects Review Board at the University of Washington to perform this study, and all patients gave written informed consent. Subjects for study consisted of 56 consecutive patients presenting for endoscopic sinus surgery performed by a single surgeon. Patients were excluded if underlying medical conditions precluded their safe participation. Patients were randomly assigned to receive propofol or isoflurane for anesthesia. Severity of sinus disease was graded based on computed tomographic scan appearance of paranasal sinuses according to the classification of Friedman et a1,2 as modified by Gliklick and Metson. Randomization was performed within each category of disease using balanced blocks of four to ensure approximately equal numbers of patients within each treatment group and with each disease category. Routine monitors included oscillometric blood pressure, electrocardiography, and pulse oximetry. Patients were pretreated with oxymetazoline (Afrin; Schering-Plough, Kenilworth, NJ) nasal spray times three and 1.0 mg of intravenous midazolam. Anesthesia was induced by intravenous alfentanil, 10 pglkg, and propofol, 2.5 mglkg, followed by maintenance of anesthesia with 60% nitrous oxide (N,O). Twenty-six patients then received isoflurane (initial end-tidal concentration, O.S%] and 30 patients received propofol (initial target plasma concentration, 4.0 p,g/mL). Propofol was delivered using a computer-controlled infusion pump (STANPUMP). All patients also received a continuous infusion of alfentanil designed to maintain a plasma concentration of approximately 35 to 41 ng/mL. Vecuronium, 0.5 mglkg, was administered intravenously for paralysis during the procedure. Patients were mechanically ventilated at 10 mL/kg, with the rate adjusted to maintain an end-tidal Pcoz of 35 mm Hg. During surgery, the patient was maintained in 10” reverse Trendelenberg to minimize venous bleeding. The initial target concentration of isoflurane or propofol was increased (up to a maximum of 1.2% isoflurane end-tidal concentration or 8 pg/mL of propofol) as required to keep mean arterial pressure (MAP) in the range of 65 to 80 mm Hg. If doubling the initial target concentration was not adequate to control blood pressure, labetolol was used, starting with 5.0 mg intravenously and doubling subsequent doses until the desired level of blood pressure was attained. Alfentanil and NzO target concentrations were kept constant. Before and during surgery, fluid was administered according to the following algorithm: 3 mL/ kg as a preoperative bolus, 8 mL/kg for the first 2 hours of surgery, and 4 mL/kg for the third hour. This fluid regimen was designed to partially replace overnight fluid deficits [assuming a deficit of 1.5 ml/kg/h for 4 hours) and to replace an anticipated blood loss of 250 mL with crystalloid on a 3:l basis. Surgery was preceded by an injection of 1%

Measurement

of Blood Loss

To measure blood loss, all shed blood was collected in suction canisters to which a known quantity of heparin had been added. At the termination of surgery, the canisters were mixed, the volume measured, and lo-mL aliquots were frozen for subsequent determination of hemoglobin (Hb) concentration. Hb concentration was measured using the cyanmethemoglobin technique in fluid from suction canisters and in blood samples obtained from patients at the beginning and end of surgery.4 The mean Hb concentration was used to calculate blood loss during surgery as follows:

TABLE 1.

Grading

System

for Assessing

Operative

Conditions Grade 1

2 3

4

Definition Minimal/no bleeding, excellent conditions Modest bleeding, modest impairment of operating conditions but quite tolerable Significant bleeding, significant impairment of operating conditions making it quite difficult to work. Change anesthetic agents Severe bleeding, operating conditions virtually intolerable, and/or consider altering original surgical plan

98

PAVLIN

Total Hb in shed blood

TABLE 2. Demographic of the Study Population

= ([Hb] in canister) (g/mL) X

(volume of canister) (mL)

Statistical

Propofol (n = 30)

Age (Y)

= (total Hh in shed blood [g]) + mean plasma Hh concentration (g/mL)

Characteristics

(1)

Volume of shed blood

(21

Analysis

Mean values of continuous variables measured once were compared by unpaired t-test. Proportions were compared by chi-squared with continuity correction or Fisher’s exact test when appropriate. Time-averaged values in which a single value was computed for each subject for each sinus were

determined for comparison of bleeding assessments for specific sinuses. A permutation test5 was used to assess the null hypothesis of no overall differ-

ence between treatment groups over time for all sinuses combined. As part of that analysis, the means of the bleeding assessment values at each

time point were computed for each treatment group. The difference of the means (mean isoflurane minus mean propofol) was calculated at each time point and the differences for all time points were combined into a single statistic by computing the mean with each time-point value weighted by the inverse of the pooled variance estimate.6 Because of the extremely large number of ways to permute the observed differences, we did not compute the difference statistic for each permutation, but chose 10,000 randomly generated computer permutations from which to compute the P presented. A p value of I .05 was considered significant throughout.

RESULTS The demographic characteristics of the patients studied are listed in Table 2. There were no significant differences in treatment groups with respect to age, weight, distribution by sex, disease category, or duration of surgery. Average intraoperative MAP was 3 mm Hg greater (P < .0167) in the propofol group. The results of bleeding assessments and measurements of blood loss are shown in Table 3 and Fig 1. There was no significant difference between groups in the measured rate of blood loss (in milliliters per kilogram per minute) or in the total blood loss (uncorrected for patient weight or surgical duration). There were, however, differences identified between the two groups with regard to bleeding assessment scores by the surgeon. When analyzed by specific sinus, mean

Height (cm) Weight (kg) Women (%) Disease category (%) 1 2 3 4 Surgical duration (min) Total time (%)/sinus Ethmoid Sphenoid Frontal Maxillary Unrecorded lntraoperative fluids (mL) Preinduction MAP (mm 4) lntraoperative (mm NOTE. percent.

lsoflurane (n = 26)

422 12 175 ? 10 84? 16 40

432 11 173 ?I 10 78% 16 57

13 80 7 0 136 ” 50

4 88 0 8 130 2 55

35 12 25 17 10 1,588 k 380

34 10 26 18 12 1,523 5 445

98t

12

99k

16

MAP 74 -c 6

W Values

ET AL

expressed

as mean

71 t4

+ standard

deviation

or

bleeding assessment scores were less in the propofol group in the ethmoid and sphenoid sinuses (P = .0475 when considered together, or P = .0297 when ethmoid sinuses were considered separately). When differences between bleeding assessment scores were compared in the two groups over time using the TABLE 3. Assessment

Measured Scores

Blood

Loss

and Bleeding

Propofol (n = 30) Total blood loss (mL) Blood loss (mUkg/min) Bleeding assessment scores Ethmoid Sphenoid Frontal Maxillary Patients (%) Mean score >2.0 Mean score >3.0 Required agent switch 04 NOTE. Values percent. *P = .04 when ethmoid alone.

expressed ethmoid

0.015

1.7 1.6 1.7 1.9

lsoflurane (n = 26)

189 + 0.01

& ? + %

0.5 0.3 0.5 0.5

0.015

147 2 0.01

2.0 1.9 1.8 1.7

2 2 + 2

0.5 0.8 0.5 0.5

P .352

,030’ .lOl’ ,797 .220

30 3

54 8

,033 ,899

13

27

,348

as mean and sphenoid

2 standard combined,

deviation

or

P = .03 for

PROPOFOL

AND

SINUS

99

SURGERY

Fig 1. Mean bleeding assessment scores over time for patients receiving propofol or isoflurane anesthesia. The groups were significantly different over time by the permutations’ test (P= .023). (H), Propofol; (0), isoflurane.

2.0 -

1.5-

p = ,023 1.04 0

50

100

TIME

permutations test to determine the probability that differences over time could have occurred by chance, the bleeding scores were significantly less in the propofol group (P = .023; Fig 1). The frequency of bleeding scores greater than 2.0 (the upper limit of what was considered “quite tolerable”) was greater in the isofluranegroup (58% v3O%;P = .0332). Although not statistically significant, there was a trend toward requesting an agent switch more often in the isoflurane group (7 of 26 v 4 of 30 patients). Switching agents was associated with improvement in bleeding assessment scores in both groups: for isoflurane to propofol, mean score declined from 2.7 2 0.3 (standard deviation) to 1.6 -+ 0.3, and for propofol to isoflurane, from 2.9 + 0.8 to 1.7 -+ 0.3. Improvement in score coincided with a significant decrease in MAP from 78 + 9 mm Hg to 65 + 2 mm Hg (P = .0185) when switching from propofol to isoflurane, but no change when switching from isoflurane to propofol (71 to 72 mm Hg). Recovery characteristics of the two groups are listed in Table 4. Overall, patients in both groups were slow to emerge from anesthesia, reflecting in part the use of deep anesthesia to control blood pressure. If discharge time is compared in only those patients who were able to be discharged successfully on the day of surgery, there is a trend toward shorter recovery in the propofol group (178 v 223 minutes, propofol v isoflurane; P = .0830). If, however, patients are included

150

200

(MINS)

in the comparison who underwent unplanned admissions for nonsurgical reasons, there were significant differences between groups in the length of stay in a recovery unit until discharge to home or to a hospital bed (183 v 243 minutes; P = 6186). Unplanned admission not directly related to surgery and considered TABLE 4. After

Comparison

of Recovery

Characteristics

Surgery Propofol

Admission category Discharged day of surgery Unplanned admission Planned admission Time (min) to Extubation Eye opening Ambulation* Actual discharge* (excluding all admissions) Actual discharget (including nonsurgical unplanned admissions) Frequency of emetic symptoms (?&) Nausea Vomiting Antiemetic treatment

30 (100) 16 (53)

lsoflurane

P

26 (100) 12 (46)

6 (20)

9 (35)

8 (27)

5 (19)

26226 29231 150 -c 55

14+9 14210 207 -c 72

.0277 .0146 .0211

207k64

253k68

.079

183263

243582

.019

50

65 23 62

17

50

.37

.79 .55

NOTE. Values expressed as number (%) or mean 2 standard deviation, unless otherwise noted. *Time to discharge and ambulation excludes patients who underwent unplanned or planned admissions. tTime to discharge, including patients who had unplanned admission for nonsurgical, possibly anesthesia-related, causes. Only time spent in a recovery unit before discharge to a hospital ward was included for patients with unplanned admission.

100

possibly related to the conduct of anesthesia occurred in eight patients overall. Included in this category were two patients in the propofol group, admitted for drowsiness; and six patients in the isoflurane group, two patients admitted for nausea and pain and nausea and four patients for pain control (alone). Seven patients had unplanned overnight admissions because of surgical complications (five patients for late time of day, two patients for observation). Four patients were in the propofol group, and three patients were in the isoflurane group.

DISCUSSION The concept that anesthesia may contribute to blood loss during surgery is not new. In the past, Stankiewicz7 reported that estimated blood loss was less (215 v 300 mL) in patients having local as opposed to general anesthesia for endoscopic sinus surgery. Comparing general anesthetic techniques, Yoshikawa et al* reported no difference in blood loss during radical maxillary sinus operations when comparing patients anesthetized by halothane, enflurane, or neuroleptanesthesia. In the present study, we observed that assessed operative conditions were, on average, better over time using propofol opposed to isoflurane for anesthesia. When the operative site was taken into account, this effect was observed most specifically in the ethmoid and sphenoid sinuses. Despite apparently superior operating conditions in the propofol group, there was no significant overall difference in the average rate of blood loss between the two groups. The apparent lack of correlation between these two observations may reflect the fact that overall, the total blood loss is relatively small and not necessarily reflective of operating conditions throughout the entire operation. A small amount of bleeding into the ethmoid sinus may impair operating conditions more than bleeding into the maxillary sinus, although ultimately, total blood loss may not be different. Also, a greater percentage of operating time was spent working in the ethmoid and sphenoid as opposed to the maxillary sinuses (Table 2), which may have accentuated differences between groups over time. The observation that bleeding in the ethmoid and sphenoid sinuses appears to be

PAVLIN

ET AL

preferentially affected by propofol is interesting in view of the fact that blood flow to the ethmoid, sphenoid, and frontal sinuses is supplied by branches of the internal carotid artery (ethmoidal arteries and supraorbital arteries), as well as branches of the sphenopalatine artery arising from the external carotid artery system. Of note, the sphenoid bone, which is particularly prone to bleeding, receives its arterial supply from branches of the internal carotid system. In contrast, the maxillary sinus receives its blood supply from branches of the external carotid artery (posterior superior alveolar, infraorbital, and anterior superior alveolar arteries).gJO It is known that propofol (as well as barbiturates and opioids) depresses cerebral blood flow in parallel with depression of cerebral metabolic rate.** In contrast, isoflurane is a cerebrovasodilator when administered at concentrations greater than 1 MAC (the minimum alveolar concentration required to prevent movement in 50% of patients).lz Thus, reduction of cerebral blood flow with propofol would be expected to decrease blood flow to the ethmoid, sphenoid, and frontal sinuses and potentially diminish bleeding related to arterial inflow. In peripheral tissues, including the maxillary sinus, this differential effect may not exist because both propofol and isoflurane are known vasodilators. The peripheral circulatory effects of propofol are mediated by cerebral depression of sympathetic tone in blood vessels, whereas isoflurane causes direct relaxation of smooth muscle in vessel walls both inside and outside (the skull).13 Thus, isoflurane may also have a greater tendency to reverse the vasoconstrictor effects of locally injected epinephrine.14 With patients in reverse Trendelenberg position, the rate of venous bleeding should be related to central venous pressure and the height of the bleeding site relative to the position and pressure of the right atrium. Thus, fluid replacement, patient state of hydration, and patient position should be the most important determinants of venous bleeding. Conceivably, propofol may preferentially minimize bleeding that is arteriolar in nature and therefore relatively rapid and difficult to control, whereas there may be no significant advantage with regard to venous bleeding. In fact, isoflurane, because it is a more potent

PROPOFOL

AND

SINUS

101

SURGERY

vasodilator, may have a greater effect in dilating capacitance vessels and decreasing systemic venous pressure. This might explain why some patients actually improved when switching from propofol to isoflurane, if bleeding was primarily venous in nature, whereas other patients appeared to benefit from switching from isoflurane to propofol. In support of this argument, improvement that occurred when switching from propofol to isoflurane was associated with a decrease in blood pressure, whereas there was no change in blood pressure when improvement occurred by switching in the reverse direction. Additionally, the decline in systemic arterial pressure associated with switching to isoflurane may also have contributed to reduced bleeding from either the internal or external carotid systems. Other options not addressed by this study include the use of alternative methods of reducing cerebral blood flow, (ie, controlled hyperventilation) or other methods of reducing arterial or venous blood pressure (ie, use of more potent vasodilators, such as nitroprusside, nitroglycerine, or fluid restriction). Of interest, Giannelli et all5 reported in abstract form that bleeding during septoplasty was less in patients who received clonidine (an CQagonist antihypertensive) before propofol or isoflurane anesthesia and concluded that the combination of propofol with clonidine was most effective in reducing surgical bleeding. The effect of isoflurane and propofol on coagulation parameters was also not addressed by this study. However, Hansen et all6 reported in abstract form that propofol, but not isoflurane, inhibited platelet function in vitro, measured by aggregometry or in vitro bleeding time, but had no effect on in vivo bleeding time or intraoperative bleeding tendency. In summary, we conclude that propofol anesthesia is on average associated with superior operating conditions compared with isoflurane for endoscopic sinus surgery. A second conclusion is that there is variation between patients and regardless of which anesthetic agent was used initially, switching to the alternate agent improved operating condi-

tions when bleeding visibility.

was seriously

impairing

ACKNOWLEDGMENT The authors wish to acknowledge that STANPUMP was provided without charge by S. Shafer, Anesthesiology Service (112A), PAVAMC, 3801 Miranda Ave, Palo Alto, CA 94304.

REFERENCES 1. Blackwell KE, Douglas AR, Kapur P, et al: Propofol for maintenance of general anesthesia: A technique to limit blood loss during endoscopic sinus surgery. Am J Otolaryngol 14:262-266,1993 2. Friedman WH, Katsantonis GP, Sivore M, et al: Computed tomography staging of the paranasal sinuses in chronic hyperplastic rhinosinusitis. Laryngoscope 100: 1161-1165,199O 3. Gliklich RE, Metson R: A comparison of sinus computed tomography (CT) staging - systems for outcome re_ search. Am JR&o1 8:291-297,1994 4. Schoen I. Solomon M: Control of blood haemoelobin determinations by a simple effective method. J Clin Fatho 15:44-46,1962 5. Cox DR, Hinkley DV (eds): Theoretical Statistics. London, England, Chapman&Hall, 1974, pp 182-187 6. Rosner BA: Fundamentals of Biostatistics [ed 4). Belmont, CA, Wadsworth, 1995, p 259 7. Stankiewicz JA: Complications in endoscopic intranasal ethmoidectomy: An update. Laryngoscope 99:686690,1989 8. Yoshikawa T, Sano K, Kan T Clinical assessment of anesthesia and estimated blood loss during maxillary sinus surgery. Anesth Prog 36:242-248,1989 9. Hollinshead WH (ed): Anatomy for Surgeons, vol 1, The Head and Neck (ed 2). New York, NY, Harper & Row, 1968, pp 287-297 10. Anson BJ (ed): Morris’ Human Anatomy [ed 12). New York, NY, McGraw Hill, 1966, p 197 11. Haberer JP, Audibert G, Saunier CG, et al: Effect of propofol and thiopentone on regional blood flow in brain and peripheral tissues during normoxia and hypoxia in the dog. Clin Physiol13:197-207,1993 12. Stoelting RK (ed): Pharmacology and Physiology in Anesthetic Practice (ed 2). Philadelphia, PA, Lippincott, 1991, pp 33-48 13. Robinsea BJ, Ebert TJ, Obrien TJ, et al: Mechanisms whereby propofol mediates vasodilation in humans. Anesthesioldgy B&64-72, 1997 14. Pavlin EG. Su TY: Cardioaulmonarv nharmacoloev. in Miller RD (ed): Anesthesii (ed 4). i\iew York, I%‘, Churchill Livingstone, 1994, pp 152-153 15. Gianelli M, Pratiu R, Barontini L, et al: Comparison of four anesthetic techniques on bleeding and adrenaline side effects during septoplasty. Br J Anaesth 76:24, A77, 1996 (suppl2, abstr) 16. Hansen E, Kutz N, Marx B, et al: Propofol inhibits platelet function. Anesthesiology 79:A428, 1993 (suppl, abstr)