Epidural anesthesia often involves the administration of large amounts of a local anesthetic solution. To avoid life-threatening central nervous system or cardiac toxicity associated with accidental intravascular injection, it is common practice to inject a test dose containing a small amount of a local anesthetic plus 10 or 15 μg epinephrine (1,2). Traditionally, heart rate (HR) increases ≥20 bpm and subjective symptoms, including circumoral pallor, palpitation, drowsiness, and tinnitus, have been regarded as positive indicators for intravascular injection of the epinephrine test dose, with varying degrees of efficacy (3,4). However, the recognition of these symptoms is obscured, and, thus, the reliability of the presence of the symptoms is decreased, by typical sedation with fentanyl and midazolam (5), which may often be administered before or during performing regional anesthesia. It is more important to note that efficacies of objective hemodynamic criteria, including HR and systolic blood pressure (SBP) changes, in sedated patients have never been addressed. In addition, the usefulness of the more contemporary T-wave criterion has never been tested in awake or sedated patients (6,7).
Accordingly, we designed this prospective, randomized study to determine and compare hemodynamic and T-wave changes to, and efficacies of, a simulated IV test dose containing 15 μg epinephrine on the basis of the HR, SBP, and T-wave criteria in patients administered midazolam alone, midazolam plus fentanyl, or no sedative.
The study protocol was approved by the Human Research Committee of the University of Akita School of Medicine, and informed consent was obtained from each patient. Sixty nonpregnant, ASA physical status I patients, free of cardiovascular disorders, who were scheduled to undergo elective surgeries were enrolled.
All patients arrived at the operating room after an 8- to 10-h fast without premedication. Standard lead II electrocardiogram and oxyhemoglobin saturation (Spo2) were monitored continuously throughout the study. A radial arterial catheter, with which subsequent blood pressure (BP) measurements were made, was placed after local anesthesia infiltration. Lactated Ringer’s solution was maintained at a constant rate of approximately 15 mL · kg−1 · h−1 through an 18-gauge IV catheter until the end of the study. All patients were subsequently randomized to one of three groups (n = 20 each) according to computed random numbers: a Midazolam group, a Midazolam Plus Fentanyl group, and a Control group. The Midazolam group patients received IV midazolam in 1-mg increments every 2 min until they were sedated; i.e., they closed their eyes without stimulus, but were easily arousable by verbal command. The Midazolam Plus Fentanyl group patients first received 2 μg/kg of IV fentanyl over 2 min, followed 3 min later by IV midazolam in 1-mg increments separated by 2 min until they were similarly sedated. The Control group patients did not receive sedatives, but were left undisturbed for 10 min. The administration of oxygen 6 L/min via a face mask was started immediately before the first dose of midazolam or fentanyl was given, but the patients remained blinded to the treatment. Then, each group of patients received normal saline 3 mL IV, followed 4 min later by 1.5% lidocaine 3 mL containing 15 μg epinephrine (1:200,000) IV as a simulated intravascular test dose via a peripheral venous line over 5 s. Continuous records (strip-chart) of HR, SBP, and lead II of the electrocardiogram were obtained after saline and the test dose injections, from which HR and SBP were analyzed at 20-s intervals for 4 min. In addition, maximum HR and SBP responses were noted. Subsequent measurements of T-wave amplitude were made at its maximum deflection and at 60-s intervals for 4 min. Hemodynamic and T-wave measurements were made separately, at random order, by an observer (TK) who remained blinded to the treatment group of patients as well as to the study drug. The high- and low-frequency filters of electrocardiography were 0.3 and 40 Hz, respectively (monitor mode). The calibration of the recorder was set at 0.5 mV/cm (double amplitude), and the chart speed was set at 25 mm/s (half-speed). All hemodynamic determinations were made with the patient in the supine position before scheduled surgeries.
A power analysis based on a previous report revealed that more than 16 patients would provide a power >0.8 (P = 0.05) for detection of a 25% difference in paired hemodynamic responses (6,8). Positive HR, SBP, and T-wave changes to the IV test dose were prospectively defined from previous reports: positive if an HR increase was ≥20 bpm (4), an SBP increase was ≥15 mm Hg (4), and a decrease in T-wave amplitude was ≥25%(6,7), occurring within 2 min of study drug administrations. We calculated sensitivity (true positives/[true positives + false negatives]), specificity (true negatives/[true negatives + false positives]), and positive (true positives/[true positives + false positives]) and negative predictive values (true negatives/[true negatives + false negatives]). All values were presented as mean ± sd and 95% confidence interval, as indicated. Statistical analysis was performed by two-way analysis of variance for repeated measurements with respect to time and groups and, when a significant difference was detected, was followed by unpaired Student’s t-test with Bonferroni’s correction. Intergroup differences in demographic data and test dose efficacies were also compared by using an unpaired Student’s t-test with Bonferroni’s correction and χ2 or Fisher’s exact probability tests as appropriate. P < 0.05 was considered to be statistically significant.
There were no significant differences in terms of patients’ age, weight, height, or sex distribution among the three groups (Table 1). Hemodynamic data and T-wave amplitudes before sedative administrations were also similar among groups. Mean doses of IV midazolam administered in the Midazolam and Midazolam Plus Fentanyl groups were 5.1 ± 1.5 and 2.6 ± 0.5 mg, respectively. Both SBP and diastolic BP decreased significantly after sedation compared with those before sedation in the Midazolam and Midazolam Plus Fentanyl groups, whereas they remained unchanged in the Control group (Table 1). T-wave amplitudes remained unchanged after sedation compared with those before sedation in the three groups. No significant difference was seen in BP, HR, and T-wave amplitude at any interval among the three groups.
IV injections of saline produced virtually no changes in HR, SBP, or T-wave amplitude in any group (Table 2). Compared with preinjection values, IV injection of the epinephrine test dose produced significant increases in HR between 40 and 120 s, 40 and 100 s, and 40 and 80 s, with their maximum changes occurring at 50 ± 9 s, 45 ± 6 s, and 48 ± 8 s in the Control, Midazolam, and Midazolam Plus Fentanyl groups, respectively (Fig. 1). The mean maximum increase in HR of the Midazolam Plus Fentanyl group was significantly less than in the Control and Midazolam groups (Table 2, P < 0.05). However, test dose injections produced significant increases in SBP between 80 and 220 s, 60 and 180 s, and 80 and 200 s, with their maximum values occurring at 79 ± 17 s, 81 ± 14 s, and 91 ± 20 s in the Control, Midazolam, and Midazolam Plus Fentanyl groups, respectively (Fig. 2). In contrast with HR, there was no significant difference in the maximum SBP increases among the three groups (Table 2). All patients who received the test dose, but none who received saline, developed decreases in T-wave amplitude by ≥25% within 60 s of injections (Fig. 3). Beat-by-beat coefficients of variations of T-wave amplitudes were <4% in all patients after they received saline. Compared with preinjection values, significant decreases in T-wave amplitudes were seen until the end of the 4-min observation period in all groups. Mean maximum percentage decreases were similar among the three groups and occurred at 38 ± 12 s, 35 ± 10 s, and 37 ± 8 s in the Control, Midazolam, and Midazolam Plus Fentanyl groups, respectively (Table 2).
In the Control and Midazolam groups, all subjects developed maximum HR increases ≥20 bpm, maximum SBP increases ≥15 mm Hg, and maximum T-wave decreases ≥25% upon test dose injections. Because no patient in these groups met the hemodynamic and the T-wave criteria after IV saline, sensitivity, specificity, and positive and negative predictive values were all 100% on the basis of the HR, SBP, and T-wave criteria (Table 3). In contrast, 6 of the 20 patients in the Midazolam Plus Fentanyl group did not reach the threshold of 20 bpm upon test dose injection, resulting in 70% sensitivity on the basis of the HR criterion (Table 3, P < 0.05 versus the Control and Midazolam groups by Fisher’s exact probability test). However, all patients in the Midazolam Plus Fentanyl group met the SBP and the T-wave criteria after IV test dose injection, and none met these criteria after IV saline, resulting in 100% efficacy on the basis of the SBP and the T-wave criteria (Table 3).
Two patients in the Control and Midazolam Plus Fentanyl groups and one in the Midazolam group developed a negative T wave, which spontaneously resolved within 2 min of the test dose administrations. No ventricular or supraventricular arrhythmia was observed in any patient throughout the study. None developed Spo2 <97%.
One major finding of our study is that the combination of midazolam and fentanyl given IV attenuated HR response to, and thus decreased the efficacy of, a simulated IV test dose containing 15 μg epinephrine compared with IV midazolam alone or no sedative. The effect of IV fentanyl, per se, on the chronotropic response to IV epinephrine has never been addressed in humans. Fentanyl affects central sympathetic and parasympathetic outflow and decreases the conductance of membrane currents via peripheral μ-receptors in the rabbit’s sinoatrial node (9,10), suggesting that fentanyl may directly or indirectly alter the chronotropic effect of epinephrine on the sinoatrial node. Indeed, our recent study (11) demonstrated, in propofol-anesthetized patients, that IV fentanyl 100 μg given at the time of general anesthesia induction with propofol decreased sensitivity from 100% to 85% for detecting intravascular injection of the identical test dose on the basis of the conventional HR criterion, compared with no fentanyl. Whether different doses of fentanyl alone or in combination with midazolam affect hemodynamic responses to and efficacies for detecting intravascular injection of the test dose remains to be determined.
On the basis of the 95% confidence intervals of the maximum HR responses after IV saline and the test dose (Table 2), the conventional HR criterion used in this study seems appropriate, but it did not ensure 100% reliability when the patient was sedated with midazolam and fentanyl. However, the T-wave criterion was 100% effective, and the peak change in its amplitude occurred earlier than that of HR, suggesting that that the T-wave morphology is a useful clinical tool and should be closely monitored immediately after the test dose is administered in such patients. The usefulness of the T-wave criterion is further underscored by the fact that its amplitude can be determined noninvasively and continuously and was reported to be more sensitive than the HR, SBP, and their combination criteria in sevoflurane-anesthetized patients when a fractional dose of the epinephrine test dose was injected IV (7). However, several factors cause the flattening and inversion of the T wave, including physical and psychological stress (12–14). Even though no false positives were observed under the circumstance of our study, ultimate validation of its reliability requires a large-scale, prospective clinical trial with real-time oscilloscopic observation, especially when the epidural blockade is wearing off during surgery.
Our study also demonstrated that the SBP criterion was applicable irrespective of the sedatives administered and confirmed the appropriateness of its testing threshold from the 95% confidence interval of absolute SBP changes. This is in contrast with a previous study, in which 1% lidocaine containing 15 μg epinephrine produced 10% false-negative responses on the basis of the same SBP criterion, with a mean maximum increase of 26 mm Hg (95% confidence interval, 18–33 mm Hg) in healthy, unmedicated volunteers (4). The considerably larger increase in SBP and its 95% confidence interval in our study may be explained, in part, by smaller patients; hence, larger doses of epinephrine per kilogram of body weight were injected in our study population, because maximum SBP increases after the simulated IV test dose are dose related in awake and anesthetized subjects (4,15). On the basis of these considerations, one may argue that the dose of epinephrine should be calculated on the basis of the weight of the subject. In contrast to the SBP changes, however, HR and T-wave alterations are not dose related (7,15). The mean maximum increments of HR in unmedicated subjects were 38 and 37 bpm after the IV test doses containing epinephrine 10 and 15 μg, respectively. Similarly, in isoflurane-anesthetized patients, IV test doses containing 15 and 22.5 μg epinephrine produced mean maximum increments in HR of 26 and 24 bpm, respectively (15). More recently, IV test doses containing 15 and 10 μg epinephrine produced mean maximum percentage decreases in T-wave amplitudes of 70% and 66%, respectively, in sevoflurane-anesthetized adults. Because, in actual clinical practice, the HR and T-wave changes are as useful indicators of intravascular injection of the epinephrine test dose as the SBP, giving the epinephrine dose on the basis of weight is yet to be justified.
The question of whether an invasive arterial catheter is mandatory to reliably detect SBP changes ≥15 mm Hg upon intravascular injection of the epinephrine test dose deserves mention. The duration of SBP variation above the testing threshold seems to last long enough to detect by use of the noninvasive BP cuff used at its most frequent measurement interval (Fig. 2). However, the inherent error for a single determination of SBP with an automated BP cuff may well exceed the threshold of the SBP criterion (16). Because placing an arterial line in all patients undergoing epidural anesthesia is neither practical nor cost-effective, whether the noninvasive BP measurements are sensitive enough to detect transient SBP changes after the IV test dose needs to be substantiated in a future study.
Our results did not elucidate the mechanism of the decreased T-wave amplitude caused by the IV test dose. Several lines of previous evidence suggest that T-wave amplitude is significantly affected by both β-adrenergic agonists and antagonists (17). In addition, epinephrine causes a reduction of serum potassium concentrations via β2 adrenoceptors (18). However, the influence of such an effect on the transient change in the T-wave amplitude is not clear, because serial changes in serum potassium concentrations were not determined in our study.
Shortcomings of our study may be threefold. First, interpretation of our results should be confined to the drugs and doses used in our study. Midazolam may be used to alleviate discomfort associated with epidural catheter placement, and fentanyl is frequently combined with midazolam to provide analgesia and augment anxiolysis. In addition, our clinical end point of the degree of sedation is the one most often practiced and had to be approximated in our protocol. However, when the tip of a multiorificed epidural catheter has migrated into a blood vessel and a fractional dose of the test dose has been injected intravascularly, as might occur in actual clinical practice, the efficacy reported here may not apply. Second, T-wave criterion used in our study requires normal T-wave morphology; i.e., preexisting T-wave abnormalities, such as in patients taking digoxin, those with left ventricular hypertrophy, or those with a history of myocardial infarction, may preclude use of the T-wave criterion (19). In addition, changes in the morphology of the electrocardiogram in other leads were not examined in our study. Finally, a continuous record of the strip-chart was subsequently analyzed by dedicated researchers, whereas the detectability of small, transient changes in T-wave amplitude on the oscilloscope was not assessed by practicing clinicians. It is more important to note that small but progressive alterations in T-wave morphology of the strip-chart may have created a bias in the determination of T-wave amplitude.
In conclusion, our results indicate that the conventional HR criterion (positive if ≥20 bpm increase) was not reliable for detecting an intravascular test dose containing 15 μg epinephrine in patients sedated with midazolam and fentanyl, whereas the SBP (positive if ≥15 mm Hg increase) and T-wave criteria (positive if ≥25% decrease in lead II) were 100% sensitive and specific in those administered midazolam, midazolam plus fentanyl, or no sedative. However, our results should be confined to young, healthy, adult subjects with normal electrocardiogram morphology. Furthermore, safety steps, such as aspiration of the epidural catheter, should always follow a positive electrocardiogram response, and a questionable catheter should be removed.
1. Liu S, Carpenter RL, Neal JM. Epidural anesthesia and analgesia: their role in postoperative outcome. Anesthesiology 1995; 82: 1474–506.
2. Mulroy MF, Norris MC, Liu SS. Safety steps for epidural injection of local anesthetics: review of the literature and recommendations. Anesth Analg 1997; 85: 1346–56.
3. Moore DC, Batra MS. The components of an effective test dose prior to epidural block. Anesthesiology 1981; 55: 693–6.
4. Guinard JP, Mulroy MF, Carpenter RL, Knopes KD. Test doses: optimal epinephrine content with and without acute beta-adrenergic blockade. Anesthesiology 1990; 73: 386–92.
5. Moore JM, Liu SS, Neal JM. Premedication with fentanyl and midazolam decreases the reliability of intravenous lidocaine test dose. Anesth Analg 1998; 86: 1015–7.
6. Tanaka M, Nishikawa T. A comparative study of hemodynamic and T-wave criteria for detecting intravascular injection of the test dose (epinephrine) in sevoflurane-anesthetized adults. Anesth Analg 1999; 89: 32–6.
7. Tanaka M, Goyagi T, Kimura T, Nishikawa T. The efficacy of hemodynamic and T wave criteria for detecting intravascular injection of epinephrine test doses in anesthetized adults: a dose-response study. Anesth Analg 2000; 91: 1196–202.
8. Fisher LD, van Belle G. Sample sizes for observational studies. In: Fisher LD, van Belle G, eds. Biostatistics: a methodology for the health science. New York: Wiley, 1993: 844–65.
9. Saeki T, Nishimura M, Sato N, et al. Electrophysiological demonstration and activation of mu-opioid receptors in the rabbit sinoatrial node. J Cardiovasc Pharmacol 1995; 26: 160–8.
10. Gautret B, Schmitt H. Multiple sites for the cardiovascular actions of fentanyl in rats. J Cardiovasc Pharmacol 1985; 7: 649–52.
11. Takahashi S, Tanaka M, Toyooka H. Fentanyl pretreatment does not impair the reliability of an epinephrine-containing test dose during propofol-nitrous oxide anesthesia. Anesth Analg 1999; 89: 743–7.
12. Guazzi M, Fiorentini C, Polese A, et al. Stress-induced and sympathetically-mediated electrocardiographic and circulatory variations in the primary hyperkinetic heart syndrome. Cardiovasc Res 1975; 9: 342–54.
13. Hijzen TH, Slangen JL. The electrocardiogram during emotional and physical stress. Int J Physiol 1985; 2: 273–9.
14. Hansen O, Johansson BW, Gullberg B. Metabolic, hemodynamic, and electrocardiographic responses to increased circulating adrenaline: effects of pretreatment with class 1 antiarrhythmics. Angiology 1991; 42: 990–1001.
15. Tanaka M, Takahashi S, Kondo T, Matsumiya N. Efficacy of simulated epidural test doses in adult patients anesthetized with isoflurane: a dose-response study. Anesth Analg 1995; 81: 987–92.
16. Borow KM, Newburger JW. Noninvasive estimation of central aortic pressure using the oscillometric method for analyzing systemic artery pulsatile blood flow: comparative study of indirect systolic, diastolic, and mean brachial artery pressure with simultaneous direct ascending aortic pressure measurements. Am Heart J 1982; 103: 879–86.
17. Contrada RJ, Krantz DS, Durel LA, et al. Effects of beta-adrenergic activity on T-wave amplitude. Psychophysiology 1989; 26: 488–92.
18. Struthers AD, Reid JL. Adrenaline causes hypokalemia in man by β2
adrenoceptor stimulation. Clin Endocrinol 1984; 20: 409–14.
© 2001 International Anesthesia Research Society
19. Verrier RL, Nearing BD. T-wave alternans as a harbinger of ischemic-induced sudden cardiac death. In: Zipes DP, Jalife J, eds. Cardiac electrophysiology: from cell to bedside. Philadelphia: WB Saunders, 1995: 467–77.