Both epinephrine 15 [micro sign]g and aspiration are commonly used to detect the IV location of an epidural catheter. Although the epinephrine test dose reliably detects intravascular catheters, (i.e., it is very sensitive), it is falsely positive in 2% of laboring women (i.e., it is nonspecific) . Although aspiration alone fails to detect 34%-81% of intravascular single holed catheters [2,3], recent studies have shown that, at most (upper limit 95% confidence interval), only 3-4:1000 aspiration-negative multiorifice catheters are in a blood vessel [4,5]. Many clinicians combine aspiration and epinephrine. When using a multiorifice catheter, aspiration effectively detects most intravascular catheters before the epinephrine test dose is injected. Because the prevalence of undetected intravascular catheters is very low, most positive responses to epinephrine will not represent intravascular catheters. We designed this study to test this assertion.
This study was approved by our human studies committee. Between August 6 and December 9, 1997, we obtained verbal consent from all parturients whose labor analgesia was directly provided or supervised by one of the authors. Women were excluded if they had a contraindication to epinephrine or if they received intrathecal sufentanil/bupivacaine. All patients were monitored with an automated blood pressure cuff and pulse oximetry. Multiorifice catheters were inserted 4-6 cm into the epidural space as part of an epidural or combined spinal-epidural (CSE) technique.
Anesthetic management was standardized. After insertion, all catheters were observed carefully and aspirated gently for cerebrospinal fluid (CSF) and blood. If aspiration revealed CSF, the catheter was either removed and replaced or used as a spinal catheter. If blood flowed freely or was continuously aspirated, the catheter was either removed and replaced or withdrawn until aspiration was negative. After negative aspiration for blood and CSF, all patients received 3 mL of 1.5% lidocaine with 5 [micro sign]g/mL epinephrine to further test for an intrathecal or IV location. Maternal blood pressure was measured before and 1-2 min after epinephrine injection. Maternal pulse was continuously monitored using a pulse oximeter probe. The diagnosis of a positive test dose response was made by clinical criteria. Rapid onset of sensory change and motor weakness suggested intrathecal injection. Tachycardia, hypertension, or maternal symptoms suggested an intravascular location. If the epinephrine test dose suggested an intravascular catheter location (i.e., was positive), the catheter was flushed and aspirated again for blood. If this repeat aspiration revealed blood, the catheter was removed and replaced. If aspiration was still negative, a dilute solution of bupivacaine and sufentanil was injected or infused into the epidural catheter. To provide epidural analgesia, we injected 10-20 mL of 0.125% bupivacaine and 10 [micro sign]g of sufentanil in 5-mL increments. Patients receiving CSE analgesia received an intrathecal injection of sufentanil 10 [micro sign]g and then were placed on an infusion of 0.083% bupivacaine with 0.33 [micro sign]g/mL sufentanil at 12 mL/h.
Twenty to thirty minutes after a bolus injection (epidural patients) or at least 2 h after beginning the infusion (CSE patients), we examined each patient for signs and symptoms of epidural analgesia. Patients with bilateral sensory change and effective labor analgesia had a positive epidural catheter. Patients with neither sensory change nor analgesia had a negative epidural catheter. Women with unilateral block, inadequate analgesia despite some sensory change, or those in the CSE group who delivered within 2 h of receiving the intrathecal drug had equivocal catheters.
Patients with negative catheters received a second epinephrine test dose. We attempted to salvage equivocal catheters in undelivered patients by withdrawing them slightly and injecting additional local anesthetic.
Data sheets were completed by the responsible anesthetist concurrently with the anesthetic. All data sheets were reviewed daily for completeness. We also reviewed the anesthetic records of all intravascular, negative, and equivocal catheters.
We provided labor analgesia for 560 women during the course of this study. Twenty-eight patients were excluded from this protocol. Twenty-one received intrathecal sufentanil/bupivacaine. In seven, the anesthetist believed that epinephrine was contraindicated (preeclampsia [n = 5], supraventricular tachycardia [n = 1], or adult respiratory distress syndrome requiring mechanical ventilation [n = 1]). Excluded patients were at more advanced cervical dilation and were more likely to receive CSE than eligible patients (Table 1).
The 532 eligible patients received 575 epidural catheters. One catheter was used in 487 women, two catheters in 44 women, 3 catheters in 5 women, and 4 catheters in one women (Table 2).
Blood was aspirated on catheter insertion 47 times. Most of these catheters were removed and replaced, but, in eight, the anesthetist withdrew the catheter until aspiration was negative. Six of these catheters were positive. One was subsequently replaced after the epinephrine test dose was positive. The other was replaced when repeat aspiration after a negative epinephrine test dose yielded some blood.
Five hundred thirty-five catheters were injected with epinephrine 15 [micro sign]g. The epinephrine test dose was negative in 525 women and positive in 10 (Table 3). Of these 535 catheters, 4 were removed and replaced (2 were IV, 2 were inadequately tested). Of the remaining 531 catheters, 515 were positive, one was negative, and 15 were equivocal. Of women with equivocal catheters, nine delivered before the catheter could be assessed. Three catheters produced one-sided blocks. Two women were rescued with additional local anesthetic, and the third was replaced. Two patients had bilateral sensory change but inadequate pain relief. Both responded to additional local anesthetic. One patient's catheter was accidentally removed before injection.
(Table 4) shows the sensitivity, specificity, positive predictive value, and negative predictive value of aspiration and epinephrine as tests to detect an intravascular location of multiorifice epidural catheters.
Veins are cannulated during 5%-15% of epidural anesthetics in parturients (2.8% in nonpregnant patients). Anesthesiologists began to be seriously concerned about unrecognized IV injection of concentrated local anesthetics in the late 1970s. In 1979, Albright  reported a series of maternal cardiac arrests and deaths in women who accidentally received IV bupivacaine. In 1981, Moore and Batra  proposed that epinephrine 15 [micro sign]g would reliably detect intravascular catheters and should be contained in a test dose to be used before the injection of large bolus doses of local anesthetic. By 1984, this epinephrine test dose had become deeply embedded in obstetric anesthetic practice .
Much of the concern about unrecognized intravascular drug injection may have related to the widespread use of single-holed epidural catheters in the United States. In the United Kingdom, where multiorifice catheters are common, bupivacaine cardiotoxicity is almost unknown. Reynolds  states, "there has been no case of intravascular catheter insertion that was not diagnosed at once on aspiration in 11,000 obstetric epidural anesthetics using three-holed catheters." In addition, in a large clinical study, we have recently shown that aspiration is a very sensitive, although not infallible, way to detect the intravascular location of multiorifice catheters .
The utility of the epinephrine test dose in obstetric labor analgesia depends, in part, on the type of epidural catheter used. Colonna-Romano and Lingaraju  used epinephrine 15 [micro sign]g to detect the intravascular location of 186 single-holed catheters that did not produce blood on aspiration. They found 14 true-positive and 8 false-positive tests. Based on these data, the positive predictive value (14/[14 + 8]) of a positive epinephrine test is 64%. Thus, only 64% of positive epinephrine tests indicate an intravascular catheter.
The situation is markedly different when using a multiorifice catheter. We have previously shown that aspiration alone detects almost all intravascular catheters . Our results suggest that, at most, only 4:1000 aspiration-negative epidural catheters would still lie within a blood vessel. In this setting, with a low prevalence of "disease" (i.e., intravascular catheters), a screening test with low specificity (i.e., epinephrine) will have a high incidence of false-positive results. We designed the current study to test this proposition by using first aspiration and then an epinephrine test dose in laboring women. As expected, most (87.5%) positive epinephrine tests did not indicate intravascular catheters. It is important to note that the diagnosis of a positive epinephrine response in this study was made exclusively by the anesthetist's assessment of the patient's clinical response (heart rate, blood pressure, subjective symptoms). Using more stringent diagnostic criteria [i.e., peak-to-peak heart rate change ] may have eliminated some of the false-positive epinephrine responses.
Clinically, one must weigh the risks of a repeat attempt to identify the epidural space against the risks of unrecognized IV injection of local anesthetics. The common risks associated with attempts to identify the epidural space include accidental dural puncture (1%-2%), intravascular catheter placement (5%-6%), and failed block (0.5%). The risks of unrecognized intravascular injection of local anesthetics depend on clinical practice. In laboring women, the only risk of unrecognized intravascular injection/infusion of appropriate doses of dilute local anesthetic/opioid solutions (e.g., <or=to0.125% bupivacaine) is failed block. Cardiac and central nervous system toxicity only occur with doses of local anesthetic that far exceed those included in a 5-mL bolus of the solutions in common use for labor analgesia.
Although this study is similar to our previous report , there are significant differences between the two designs. In our first study, patients were enrolled at the discretion of the responsible anesthetist. Thus, some patients may have been excluded because of a perceived increased risk of intravascular catheter location. In the current study, we enrolled consecutive patients to minimize the risk of bias. The original study included patients who received intrathecal local anesthetic, but because residual sensory block may cloud the diagnosis of a positive epidural catheter (bilateral sensory change and analgesia >or=to2 h after intrathecal injection), we chose to exclude those women from the current study. Finally, and most important, the original study did not include the use of an epinephrine test dose after a negative aspiration.
Two remaining issues deserve mention: attempting to withdraw an intravascular catheter until it lies within the epidural space and the possibility of multicompartment catheter placement. We previously reported that, although aspiration detects almost all multiorifice catheters that are initially inserted into a blood vessel, it is less sensitive after such a catheter has been withdrawn in an attempt to site it in the epidural space . In this study, eight intravascular catheters were withdrawn until blood no longer returned. At least one of these eight catheters was still in a blood vessel. It may be more prudent simply to replace intravascular catheters instead of trying to reposition them.
The three holes of a multiorifice catheter may lie in more than one compartment. Beck et al.  radiologically verified the catheter position in 113 patients who had received epidural anesthesia for lower extremity surgery. Thirteen (12%) of the catheters were epidurovascular. One of these patients received, at separate times, epinephrine 10 [micro sign]g via a peripheral IV line and the epidural catheter. The peripheral venous injection produced tachycardia and hypertension; the catheter injection did not. Three of the epidurovascular catheters provided surgical anesthesia.
At least two catheters in this current study may have been only partially intravascular. Both produced adequate analgesia for several hours before failing. Blood was then noted on manipulation or reinjection. In a multiorifice catheter, the port through which injectate exits depends on the speed of injection. Slowly injected fluid exits only through the most proximal hole, whereas a rapid injection exits from all ports . Thus, if the distal port of a catheter is in a blood vessel but the proximal port lies within the epidural space, the slow injection or infusion of local anesthetics will produce analgesia. Should the proximal port become blocked, analgesia will wane.
These data have several significant limitations. They apply only to laboring women. Periodic pain-related heart rate changes are responsible for much of epinephrine's poor specificity in this population . The epinephrine test dose may be more useful in nonpregnant or nonlaboring patients . These results apply only to the brand of multiorifice catheters that we used (B. Braun Medical Inc., Bethleham, PA). Although there is no a priori reason to believe that other multiorifice catheters would behave differently, such data are not currently available. Finally, these data assume an otherwise safe clinical practice when providing labor epidural analgesia (i.e., using an incremental injection and diluted local anesthetic/opioid solutions. Catheters that produced neither sensory change nor analgesia were removed and replaced, not injected with larger doses of the local anesthetic.
In summary, in this study, we confirmed that aspiration alone detects almost all (>or=to99.5%) intravascular multiorifice catheters in laboring women. In laboring women, tachycardia after an injection of epinephrine 15 [micro sign]g into a catheter that did not yield blood on aspiration most often does not indicate an intravascular catheter. We recommend that practitioners consider abandoning the routine use of an epinephrine test dose when providing epidural labor analgesia with a multiorifice catheter and that they dilute local anesthetic/opioid solutions.
1. Colonna-Romano P, Lingaraju N. Tests to evaluate intravenous placement of epidural catheters in laboring women: a prospective clinical study. Anesth Analg 1998;86:985-8.
2. Kenepp NB, Gutsche BB. Inadvertent intravascular injections during lumbar epidural analgesia. Anesthesiology 1981;54:172.
3. Leighton BL, Norris MC, DeSimone CA, et al. The air test as a clinically useful indicator of intravenously placed epidural catheters. Anesthesiology 1990;73:610-3.
4. Norris MC, Fogel ST, Dalman H, et al. Labor epidural analgesia without an intravascular "test dose." Anesthesiology 1998;88:1495-501.
5. Topkis W, Lewin S, Leighton BL, et al. Effectiveness of the Doppler air test in diagnosing intravascular multi-orifice epidural catheters [abstract]. Anesthesiology 1998;A30.
6. Albright GA. Cardiac arrest following regional anesthesia with etidocaine or bupivacaine. Anesthesiology 1979;51:285-7.
7. Moore DC, Batra MS. The components of an effective test dose prior to epidural block. Anesthesiology 1981;55:693-6.
8. Albright GA. Epinephrine should be used with the therapeutic dose of bupivacaine in obstetrics. Anesthesiology 1984;61:217-8.
9. Reynolds F. Epidural catheter migration during labour [letter]. Anaesthesia 1988;43:69.
10. Leighton BL, Norris MC, Sosis M, et al. Limitations of epinephrine as a marker of intravascular injection in laboring women. Anesthesiology 1987;66:688-91.
11. Beck H, Brassow F, Doehn M, et al. Epidural catheters of the multi-orifice type: dangers and complications. Acta Anaesthesiol Scand 1986;30:549-55.
12. Cartwright PD, McCarroll SM, Antzaka C. Maternal heart rate changes with a plain epidural test dose. Anesthesiology 1986; 65:226-8.
13. Mulroy MF, Norris MC, Liu SS. Safety steps in epidural injection of local anesthetics: review of the literature and recommendations. Anesth Analg 1997;85:1346-56.