Secondary Logo

Journal Logo

Positive Intravascular Test Dose Criteria in Children During Total Intravenous Anesthesia with Propofol and Remifentanil Are Different than During Inhaled Anesthesia

Polaner, David M., MD, FAAP; Zuk, Jeannie, PhD, RN; Luong, Kristi, MD; Pan, Zhaoxing, PhD

Section Editor(s): Davis, Peter J.

doi: 10.1213/ANE.0b013e3181c5f2dc
Pediatric Anesthesiology: Research Reports

BACKGROUND: The use of local anesthetic test doses is standard practice when performing regional anesthesia. When an intravascular test dose is administered during inhaled anesthesia, the heart rate does not increase in about 25% of children; altered T-wave amplitude is a better indicator. No studies have examined the criteria for a positive result during total IV anesthesia (TIVA) in children.

METHODS: We studied the effect of a simulated positive test dose on heart rate, arterial blood pressure, and T-wave amplitude in 17 ASA physical status I or II children receiving TIVA with propofol and remifentanil. Bupivacaine 0.25% 0.1 mL/kg with epinephrine 1:200,000 was injected IV, and vital signs and electrocardiogram were continuously monitored. Increases of heart rate and arterial blood pressure >10% and T-wave amplitude >25% of baseline were considered clinically significant changes.

RESULTS: All subjects had increased systolic and diastolic blood pressure (30.3% ± 11.7% and 49.3% ± 16.7%), which peaked within 120 s. Heart rate increases >10% of baseline occurred in 73% of subjects. T-wave amplitude increased in 33.3%, was unchanged in 25%, and decreased in 41.7% of subjects.

CONCLUSIONS: A positive test dose during TIVA is best detected by increased arterial blood pressure. Twenty-seven percent of intravascular injections were missed using heart rate criteria. T-wave amplitude is not a reliable indicator of intravascular injection during TIVA. This is in marked distinction to what is seen during inhaled anesthesia.

Published ahead of print November 21, 2009

From the Department of Anesthesia, The Children's Hospital, Denver and University of Colorado School of Medicine, Aurora, Colorado.

Accepted for publication August 31, 2009.

Published ahead of print November 21, 2009

Supported by Department of Anesthesia, The Children's Hospital, Denver, CO.

DMP is the principal investigator and was involved in the design and conduct of the study, data collection and analysis, and manuscript preparation. JZ helped in the design of the study, data collection and analysis, and manuscript preparation. KL helped with the conduct of the study, data collection, and analysis. ZP participated in the data analysis for the study.

Preliminary data from this study were presented at the Annual Meeting of the American Society of Anesthesiologists, October 26, 2004, Las Vegas, NV.

Kristi Luong is currently at Anesthesia Consultants, Denver, CO.

Address correspondence and reprint requests to David M. Polaner, MD, FAAP, Department of Anesthesia, The Children's Hospital, 13123 East 16th Ave., B090, Aurora, CO 80045. Address e-mail to

For >20 yr, the use of test doses of local anesthetics has been a standard practice when performing regional anesthesia. The technique was first described by Moore and Batra1 in awake adults. They determined that an IV injection of local anesthetic containing 15 μg of epinephrine resulted in a predictable increase of heart rate (mean increase of 31 bpm) that occurred within 23 s of injection and lasted approximately 30 s. This simulated the inadvertent IV injection of an epidural test dose and suggested that such test doses could be used to detect intravascular misplacement of an epidural catheter. Despite the widespread acceptance of this technique, no studies of test doses in either pediatric patients or in patients under general anesthesia were performed until 1990, when Desparmet et al.2 used the same methodology in children receiving halothane. They found that IV injection of a test dose did not produce an increase of heart rate in 27% of the subjects, clearly an unacceptably high false-negative rate.

Subsequent studies by other investigators have confirmed the findings of Desparmet et al. and have shown similar results when using other inhalation anesthetics.3–5 Careful observation of the ST segment of the electrocardiogram (ECG) has been shown to detect IV injection with a higher degree of reliability than either heart rate changes or arterial blood pressure changes in children under general inhaled anesthesia.6,7

The use of total IV anesthesia (TIVA) in both adults and children has increased in recent years with the introduction of new short-acting IV drugs such as propofol and remifentanil. Although a long- or intermediate-acting opioid or nonsteroidal drug can be administered at the end of the procedure for postoperative analgesia, regional blockade, administered while the child is anesthetized, is often used for postoperative analgesia in children and has numerous advantages, including fewer undesirable side effects and a long duration of action. There are no published data on the efficacy of test doses to detect inadvertent intravascular injection of local anesthetic during TIVA in children. We studied the effect of an intravascular injection of bupivacaine with epinephrine on heart rate, T-wave amplitude, and arterial blood pressure to determine the reliability and the best indicator of a positive test dose during TIVA. We used the same methodology used in the investigations described above to determine whether standard test doses are effective at detecting an intravascular injection of local anesthetic.

Back to Top | Article Outline


After local IRB approval, 17 ASA physical status I or II children, aged 8 mo to 12 yr, were enrolled in the study. Informed consent was obtained from the children's parents, and assent was obtained from children older than 7 yr. All children were scheduled to undergo elective surgical procedures under general anesthesia and were deemed suitable for TIVA with propofol (Baxter, New Providence, NJ) and remifentanil (Abbott Laboratories, Abbott Park, IL) by the attending anesthesiologist responsible for their clinical care. Exclusion criteria were factors that might increase the risk of toxicity for local anesthetics and included age <6 mo, a history of cardiac disease or dysrhythmias, or a history of seizures. A premedication of oral midazolam (Roxane Laboratories, Columbus, OH) 0.3 mg/kg was given at the discretion of the attending anesthesiologist. Anesthesia was induced by inhalation of sevoflurane (Abbott Laboratories) in 40% oxygen/60% N2O or, in 1 subject in whom IV access was in place, with 3.5 mg/kg of propofol. As soon as consciousness was lost, an IV cannula was placed, the sevoflurane was discontinued, and TIVA with remifentanil and propofol was begun. The fresh gas flow was increased to at least 10 L/min to speed the washout of sevoflurane. Ten milliliters per kilogram of lactated Ringer's solution was rapidly administered. No neuromuscular blocking drugs, atropine, or other drugs were given. An end-tidal CO2 value in the normal range was achieved either with spontaneous ventilation or pressure-limited ventilation.

Remifentanil, 100 μg/mL, was freshly mixed in propofol at a concentration of 20 μg of remifentanil per milliliter (10 mg) of propofol. The infusion was begun at 100 μg · kg−1 · min−1 of propofol/0.2 μg · kg−1 · min−1 of remifentanil and controlled with a Medfusion 2010 syringe pump (Medex, Dublin, OH). The infusion was run for 10 min before the initiation of the study to allow for adequate washout of the sevoflurane and for stabilization of the TIVA.

After at least 10 min, when the end-tidal concentration of sevoflurane reached 0%, control values of heart rate, arterial blood pressure, and ECG were obtained. A simulated positive test dose of 0.1 mL/kg (3 mL maximum) of 0.25% bupivacaine with epinephrine (AstraZeneca, Wilmington, DE), 1:200,000 (0.25 mg/kg of bupivacaine and 0.5 μg/kg of epinephrine), was administered through the IV line at time 0. ECG (lead II), heart rate, and pulse oximetry were continuously monitored and recorded. Arterial blood pressure was measured by automated oscillometric cuff every 60 s. ECG was recorded on either a paper strip chart recorder (Subjects 1–12) or using a computerized data acquisition system, which digitized the ECG signal directly to a hard drive (Datex-Ohmeda s/5 Data Collect, Helsinki, Finland). Data were collected until all values returned to baseline (<3 min in all subjects). No other medications were administered until after the end of the study, and no surgical stimulation occurred during the study period.

The criteria set for positive test dose detection were as follows:

  • Heart rate: increase of >10% from baseline,
  • T-wave amplitude: change of >25% from baseline,
  • Arterial blood pressure: increase of systolic or diastolic blood pressure of >10% from baseline.

These values were chosen to reflect the smallest change that a clinician would be likely to easily detect by observation during clinical care and are consistent with criteria used in previous investigations of test doses during general anesthesia.

T-wave amplitude data were measured by hand from paper ECG strips or, when available digitally, using the measurement tools in the s/5 Data Collect program. The hand-measured data were measured in millimeters, and the digitally collected data in millivolts. The height of the T-wave was measured as the amount of deflection from the ST segment. The maximum change in T-wave amplitude (negative or positive) was measured after review of the entire strip chart or digital recording to identify the complex with the greatest change from the baseline value. T-wave change data are expressed in percent change from baseline.

Data were analyzed with SAS for Windows (SAS Institute, Cary, NC). Heart rate changes were evaluated with a signed rank test. A quadratic regression of the heart rate over the 60-s window after test dose injection was performed. Arterial blood pressure data were evaluated by a sensitivity analysis using a binomial proportion to determine confidence limits and by signed rank test. T-wave data were analyzed with 2-tailed t-tests, and upper and lower confidence intervals were determined.

Back to Top | Article Outline


Seventeen subjects were enrolled in the study (Table 1). Two subjects were eliminated from the analysis, one because of data recording malfunctions and the other because it was discovered after administration of the test dose that the left limb and arm ECG leads were inadvertently reversed. No adverse events occurred in any subject. Although the original intent was to enroll 60 subjects, a number based on power analysis of previous studies using this methodology, the IRB, at their annual review of the study, voiced concerns that the small transient arterial blood pressure increases placed our subjects at increased risk of adverse events. At this point, an interim analysis was performed, which showed that statistical significance had already been reached even with this small number of subjects, and enrollment in the study was concluded.

Table 1

Table 1

Back to Top | Article Outline

Heart Rate

Positive heart rate criteria (an increase of >10% from baseline) were met in only 73% of subjects, although an increase of some magnitude was seen in every subject (Fig. 1). Heart rate increased by a mean of 19.5% (95% confidence interval 14.5%–28.5%, P < 0.0001). In those whose heart rates increased >10%, that threshold was detected within 40 s of test dose injection (mean peak, 33 s) and returned to baseline by 60 s (Fig. 2).

Figure 1.

Figure 1.

Figure 2.

Figure 2.

Back to Top | Article Outline

T-Wave Amplitude

T-wave amplitude increased in 33.3%, was unchanged in 25%, and decreased in 41.7% of subjects (P = 0.909 by paired t-test) (Fig. 3).

Figure 3.

Figure 3.

Back to Top | Article Outline

Arterial Blood Pressure

Systolic blood pressure increased by a mean of 26% ± 12.5% (95% confidence interval 19.5%–32.0%, P < 0.0001) and diastolic by 40.5% ± 21.1% (95% confidence interval 29.8%–51.1%, P < 0.0001) of baseline values (Fig. 4). Diastolic blood pressure increased above the 10% threshold in all subjects studied, and systolic blood pressure did so in all but 1 subject, whose pressure increased by 9%. When heart rate did increase, it always preceded the increase in blood pressure; however, without continuous measurement of arterial blood pressure, we cannot determine this with surety. Increases in blood pressure always peaked within 120 s of test dose injection and returned to near baseline values within 180 s.

Figure 4.

Figure 4.

Back to Top | Article Outline


Previous reports of test doses administered during general inhaled anesthesia have noted that relying on heart rate changes to detect intravascular injection carries a false-negative rate of approximately 25%. Our data obtained during TIVA with propofol and remifentanil are similar. Changes in T-wave amplitude have been shown to be a reliable indicator during inhaled anesthesia, detecting 100% of IV injections during sevoflurane anesthesia and 90%–95% during halothane anesthesia.6 In our subjects, however, T-wave alterations were a completely unreliable and variable sign during TIVA. This also differs from an adult study during propofol and N2O anesthesia, whereby subjects consistently had decreases in T-wave amplitude when a simulated positive test dose was administered.8

In our study, only arterial blood pressure was a reliable indicator of IV injection during propofol-remifentanil TIVA in children. The diastolic blood pressure was a particularly sensitive indicator, increasing >10% in every subject. A ≥10% increase over baseline systolic blood pressure was measured in 87% of our subjects, and all subjects evidenced a >9% increase. Oscillometric blood pressure measurements, which measure mean arterial blood pressure (the pressure at which oscillometric amplitude is greatest) and extrapolate the systolic and diastolic blood pressure using a computerized algorithm, may be inaccurate up to ±5 mm Hg as per the 1992 AAMI SP-10 standard but have been found to be closely correlated with both auscultated and intraarterial pressures under most circumstances.9,10 We obtained arterial blood pressure measurements at 1-min intervals and thereby may have missed the true peak pressure or overestimated the time to arterial blood pressure change that is seen after the injection of the test dose.

In our hospital and many other institutions, bupivacaine remains the most commonly used local anesthetic for regional blockade in children older than 6 mo of age. It is often our practice to use bupivacaine for the test dose, although lidocaine with epinephrine is also sometimes administered. We recommend some caution at extrapolating these results to a lidocaine and epinephrine test dose because these drugs might have different effects during TIVA, although the effects reported in studies with inhaled anesthesia are similar with both local anesthetics.

The reason for the inconsistent T-wave responses is not clear. Propofol has been shown to have no effect on atrioventricular node and sinuatrial nodal conduction in electrophysiologic studies in humans.11 It has been demonstrated to decrease the dose of epinephrine required to induce dysrhythmias in dogs.12 The effect of propofol on the Q-T interval and on the Q-T interval corrected for heart rate (Q-Tc) is controversial, although most investigators have found no change.13–15 It can increase the His-ventricular interval in the pig, but has no effect on repolarization.16 Remifentanil increases vagal tone, resulting in bradycardia; we are not aware of any data regarding its effect on cardiac conduction.

Definitive conclusions from our data are limited by the small sample size because of the premature termination of enrollment. Nevertheless, statistical analysis showed that our results reached significance. Furthermore, the magnitude of changes measured, and, in the case of the T-wave data, the distribution of positive, negative, and no changes was easily detected clinically, suggesting that the results are applicable to the clinical setting. Although we did not enroll a control group or use a sham injection (e.g., administration of a saline bolus rather than a simulated test dose), the baseline period before any drug injection allowed each subject to serve as his or her own control. There was no stimulation or alteration in anesthetic depth during the baseline or study periods, thus no reason to suppose that physiologic changes measured were due to anything other than the administration of the simulated test dose. The previous studies using this methodology have all shown that false-positive findings do not occur under these conditions.

Back to Top | Article Outline


Our study found that, in marked distinction to what is seen during inhaled anesthesia, T-wave morphology is highly unreliable as an indicator of intravascular injection of bupivacaine with epinephrine during TIVA with propofol and remifentanil in children. Increases in arterial blood pressure, particularly diastolic blood pressure, seem to be the most reliable positive test dose criteria during TIVA under the conditions studied. Heart rate seems to have similar reliability during TIVA as during inhaled anesthesia and is inadequately sensitive to detect >75% of intravascular injections.

Despite the small sample size of this study, these data strongly suggest that the usual positive test dose criteria are inadequate to detect intravascular injection during TIVA and should not be relied upon. We recommend measuring arterial blood pressure at frequent intervals of no longer than 1 min after injection of the test dose and using the diastolic blood pressure as the most reliable criterion of an intravascular injection during TIVA. Although less potentially toxic levo-enantiomers of local anesthetics such as ropivacaine are becoming more frequently used, test dosing will continue to be necessary to detect intravascular misplacement of injection of even these safer drugs because they are not without risk of toxicity from intravascular injection.17,18 It is obvious that any test dose regimen is potentially fallible, and that our data are based on small numbers of subjects; therefore, incremental dosing is always advised. We look forward to further studies with larger numbers of subjects to corroborate our results.

Back to Top | Article Outline


The authors thank Jane Gralla, PhD, for her contributions to the statistical analysis, and Zachary Desmond, BSEE, and James Carollo, PhD, PE, of the Center for Gait and Movement Analysis at The Children's Hospital for their invaluable bioengineering assistance with the digital data acquisition. They also thank Tom Ritchie and Datex-Ohmeda, Madison, WI, for their loan of an s/5 monitor for testing purposes.

Back to Top | Article Outline


1. Moore DC, Batra MS. The components of an effective test dose prior to epidural block. Anesthesiology 1981;55:693–6
2. Desparmet J, Mateo J, Ecoffey C, Mazoit X. Efficacy of an epidural test dose in children anesthetized with halothane. Anesthesiology 1990;72:249–51
3. Tanaka M, Nishikawa T. Simulation of an epidural test dose with intravenous epinephrine in sevoflurane-anesthetized children. Anesth Analg 1998;86:952–7
4. Tanaka M, Nishikawa T. The efficacy of a simulated intravascular test dose in sevoflurane-anesthetized children: a dose-response study. Anesth Analg 1999;89:632–7
5. Liu SS. Hemodynamic responses to an epinephrine test dose in adults during epidural or combined epidural-general anesthesia. Anesth Analg 1996;83:97–101
6. Kozek-Langenecker SA, Marhofer P, Jonas K, Macik T, Urak G, Semsroth M. Cardiovascular criteria for epidural test dosing in sevoflurane- and halothane-anesthetized children. Anesth Analg 2000;90:579–83
7. Fisher QA, Shaffner DH, Yaster M. Detection of intravascular injection of regional anaesthetics in children. Can J Anaesth 1997;44:592–8
8. Takahashi S, Tanaka M, Toyooka H. The efficacy of hemodynamic and T-wave criteria for detecting intravascular injection of epinephrine test dose in propofol-anesthetized adults. Anesth Analg 2002;94:717–22
9. Ling J, Ohara Y, Orime Y, Noon GP, Takatani S. Clinical evaluation of the oscillometric blood pressure monitor in adults and children based on the 1992 AAMI SP-10 standards. J Clin Monit 1995;11:123–30
10. Colan SD, Fujii A, Borow KM, MacPherson D, Sanders SP. Noninvasive determination of systolic, diastolic and end-systolic blood pressure in neonates, infants and young children: comparison with central aortic pressure measurements. Am J Cardiol 1983;52:867–70
11. Lavoie J, Walsh EP, Burrows FA, Laussen P, Lulu JA, Hansen DD. Effects of propofol or isoflurane anesthesia on cardiac conduction in children undergoing radiofrequency catheter ablation for tachydysrhythmias. Anesthesiology 1995;82:884–7
12. Kulier AH, Turner LA, Vodanovic S, Contney S, Lathrop DA, Bosnjak ZJ. Multiple agents potentiate alpha1-adrenoceptor-induced conduction depression in canine cardiac purkinje fibers. Anesthesiology 2000;92:1713–21
13. Kleinsasser A, Loeckinger A, Lindner KH, Keller C, Boehler M, Puehringer F. Reversing sevoflurane-associated Q-Tc prolongation by changing to propofol. Anaesthesia 2001;56:248–50
14. Whyte SD, Booker PD, Buckley DG. The effects of propofol and sevoflurane on the QT interval and transmural dispersion of repolarization in children. Anesth Analg 2005;100:71–7
15. Saarnivaara L, Hiller A, Oikkonen M. QT interval, heart rate and arterial pressures using propofol, thiopentone or methohexitone for induction of anaesthesia in children. Acta Anaesthesiol Scand 1993;37:419–23
16. Pires LA, Huang SK, Wagshal AB, Kulkarni RS. Electrophysiological effects of propofol on the normal cardiac conduction system. Cardiology 1996;87:319–24
17. Mather LE, Copeland SE, Ladd LA. Acute toxicity of local anesthetics: underlying pharmacokinetic and pharmacodynamic concepts. Reg Anesth Pain Med 2005;30:553–66
18. Petitjeans F, Mion G, Puidupin M, Tourtier JP, Hutson C, Saissy JM. Tachycardia and convulsions induced by accidental intravascular ropivacaine injection during sciatic block. Acta Anaesthesiol Scand 2002;46:616–7
© 2010 International Anesthesia Research Society