This article is featured in “This Month in Anesthesiology.” Please see this issue of Anesthesiology, page 5A.
ASPIRATION is the simplest of the many ways to detect intravenously positioned catheters during epidural anesthesia. However, aspiration fails to detect 33–67% of intravenous single‐orifice (open tip) catheters. [1,2]
In contrast, some investigators claim that aspiration always detects intravenous multiple‐orifice catheters. [3,4]
This study is the first large, prospective, systematic evaluation of the ability of aspiration to detect the intravenous location of multiple‐orifice catheters during labor epidural analgesia.
Materials and Methods
The Human Studies Committee of Washington University in St. Louis approved this protocol. This study involved collecting data about a standardized, accepted clinical practice. As such, it involved no additional risk to the patients. Therefore, the Human Studies Committee waived the requirement for obtaining informed consent in accordance with Title 45 of the Code of Federal Regulations, Part 46‐Protection of Human Subjects. 
One thousand twenty‐nine women (of 1,624 eligible patients) who requested neuraxial labor analgesia participated in this clinical protocol. The decision to enroll a patient was left to the individual anesthetist (staff anesthesiologist, certified registered nurse anesthetist, or anesthesiology resident). Patients received either epidural or combined spinal epidural (CSE) analgesia at the discretion of the anesthetist. We identified the epidural space using loss of resistance to air or saline. In the patients receiving CSE, we used the needle‐through‐needle technique and gave an intrathecal injection of 5 ‐ 10 [micro sign]g sufentanil via a 27‐gauge Whitacre needle. If the patient was in advanced labor, we added 2 to 2.5 mg bupivacaine to the intrathecal injection. All patients had a 20‐gauge, closed‐tip (multiple‐orifice) polyamide catheter (B. Braun Medical, Bethlehem, PA) inserted 2 ‐ 8 cm into the epidural space.
The initial testing of all catheters was performed according to a defined protocol. Thereafter, management was left to the discretion of the individual anesthetist. According to the protocol, catheters were first observed carefully and aspirated gently for return of blood or cerebrospinal fluid (CSF). If results of observation and aspiration were negative, we injected 2 ml local anesthetic (0.25% bupivacaine or 0.2% ropivacaine) to rule out intrathecal catheter location. Then, in patients receiving epidural analgesia, we injected 10 ‐ 15 ml local anesthetic with or without opioid in divided doses. In patients receiving CSE analgesia, we usually performed the intrathecal test immediately after catheter insertion and then began an infusion of 0.083% bupivacaine with 0.33 [micro sign]g/ml sufentanil (or 1 [micro sign]g/ml fentanyl) at 10 ‐ 15 ml/h.
Epidural catheters were tested within 30 ‐ 45 min of induction or, in the CSE patients, >2 h after starting the epidural infusion. We asked patients if they had obtained adequate pain relief and looked for evidence of sensory change (loss of ability to perceive cold). We defined a “positive” epidural catheter as the presence of bilateral sensory change and effective labor analgesia. If a patient had no sensory change (a “negative” epidural catheter), we checked for intravascular catheter location using the Doppler test, 
injection of 15 [micro sign]g epinephrine, or repeated aspiration of the catheter. Women with unilateral block, inadequate analgesia despite some sensory change, or those in the CSE group who delivered within 2 h of intrathecal drug injection had “equivocal” epidural catheters.
Catheters yielding either frank blood or CSF were managed at the clinician's discretion. Most were removed and reinserted. Some clinicians tried to withdraw intravenous catheters until aspiration yielded no blood. Testing and dosing of these catheters was left to individual choice. Some subarachnoid catheters were used to provide continuous spinal analgesia. If blood or blood‐tinged fluid appeared in the catheter but did not flow freely or could not be aspirated, the catheter was flushed with normal saline and aspiration was repeated.
A separate data form was completed for each epidural catheter (Figure 1
). On it, we recorded demographic data, the depth of catheter insertion, the anesthetic technique (epidural or CSE), the presence of blood or CSF in the catheter, the results of the intrathecal local anesthetic “test dose,” and the status of the epidural catheter (positive, negative, or equivocal). The data sheet also had room to record explanations of equivocal catheters and any other pertinent events. Data sheets were included in the packet of records completed for each labor anesthetic. We checked for compliance with the protocol (specifically use of aspiration alone to detect intravascular catheters, use of the plain local anesthetic test dose, and documentation of bilateral sensory change and analgesia to indicate a positive epidural catheter) by randomly reviewing 1% of the anesthetic records. Data sheets were completed by the responsible anesthetist concurrently with the anesthetic. The first author reviewed all data sheets for completeness and also reviewed the anesthetic records of all intravascular and negative catheters.
Frequency of events was compared using the chi‐squared test. Confidence intervals were calculated using standard formulas.
One thousand twenty‐nine patients (Table 1
) had 1,085 catheters inserted (Table 2
). One catheter was inserted in 977 patients, two catheters in 47 patients, 3 catheters in 4 patients, and 4 catheters in 1 patient. Catheters were replaced because they were intrathecal (n = 2), intravascular (n = 55), nonfunctional (n = 2), or had a positive response to 10 [micro sign]g epinephrine (n = 1, explained subsequently). Of the 47 patients who received two catheters, 43 were replaced because the original was intravenous, 2 because the original was intrathecal, and 2 because the original failed. The women who received three and four catheters had repeated intravascular catheters. No patient suffered any complications related to unrecognized intravenous injection of local anesthetics. Enrolled patients were more likely to receive CSE than epidural analgesia compared with the eligible patients (73% vs. 68%; P < 0.05 by the chi‐squared test).
Aspiration yielded CSF in four catheters. Three were inserted as a part of an intended CSE technique and one during attempted epidural analgesia. Two were replaced and two were used as spinal catheters. Blood appeared either spontaneously or with aspiration in 60 catheters (5.5%). Fifty‐three of these catheters were replaced immediately. Six were withdrawn until aspiration was negative. Four of the these yielded positive epidural catheters. The other two yielded blood after infusion and bolus injections of bupivacaine failed to provide analgesia or sensory change. These two catheters were then replaced. One patient declined reinsertion of an intravascular catheter. Anesthetic technique (epidural vs. CSE) did not alter the chance of aspirating either CSF or blood (chi‐squared test, P = NS).
We tested 1,023 catheters for intrathecal location using a 2‐ml bolus of local anesthetic. There were no positive responses. Sixty‐one catheters were not tested for intrathecal location. Most were removed after they yielded blood on aspiration (n = 54). Four were recognized intrathecal catheters, and three patients delivered before their catheters were tested.
We infused or injected local anesthetic in 1,023 catheters (Table 2
). There were 957 positive epidural catheters and 58 equivocal catheters (Table 3
). Fifty‐three of the equivocal catheters were inserted as part of a CSE technique. There were eight negative catheters (Table 4
). Two of these catheters were intravascular (explained previously). The other six (all inserted for epidural analgesia) were neither epidural nor intravascular.
Four catheters were not tested adequately. Three were never injected. The fourth was withdrawn after a positive heart rate response to 10 [micro sign]g epinephrine.
Two patients deserve further mention. One epidural catheter appears to have migrated into a blood vessel. This patient received 10 [micro sign]g intrathecal sufentanil as part of a CSE technique. After aspiration for blood and CSF were negative, we began an epidural infusion of bupivacaine and sufentanil. After 3 h of effective analgesia, the patient began to report pain. Aspiration revealed blood, the catheter was removed, and the patient delivered shortly thereafter with bilateral sensory change in the distribution of the sacral nerves. We classified this catheter as positive. The second patient had four epidural catheters inserted. The first two yielded blood on aspiration and had positive heart rate responses to 15 and 10 [micro sign]g epinephrine, respectively. Aspiration was negative with the third catheter, but 10 [micro sign]g epinephrine produced maternal tachycardia (degree unspecified). The catheter was removed without further testing. We classified this third catheter as “not adequately tested.” On the fourth attempt, both aspiration and epinephrine were negative. This fourth catheter proved to be a positive epidural catheter. The classification of these two catheters does not significantly change the results of this study.
These results suggest that labor epidural analgesia can be provided without the need for traditional intravenous test doses. Further study is needed before these data can be extrapolated to the use of concentrated local anesthetics for surgical epidural anesthesia.
Veins are cannulated during 5 ‐ 15% of epidural anesthetics in parturients (and in 2.8% of nonpregnant patients). Unrecognized intravascular injection of local anesthetic may be catastrophic. In the United States, the most common tests to detect intravenous catheters in parturients are subjective symptoms, vital sign changes after epinephrine injection, or Doppler detection of intravenous air. [6,7]
These test doses have become deeply embedded in anesthetic practice, yet few data document their safety, efficacy, or necessity. In contrast, aspiration alone is often used in the United Kingdom. 
Intravenous injection of both opioids and local anesthetics may produce subjective symptoms. The few systematic studies of these test doses have each included fewer than 100 women. [8–11]
Even with these small groups, false‐positive and false‐negative responses occur. Thus inquiring about subjective symptoms has limited usefulness in current obstetric anesthesia practice.
Epinephrine at a dose of 15 [micro sign]g is the most commonly debated test dose in obstetrics. Investigators doing small studies have questioned its efficacy and safety in laboring women. [12–14]
Others have argued for its routine use in obstetrics. [15–17]
Recently, Colonna‐Romano et al. 
reported their experience with 15 [micro sign]g epinephrine in 209 laboring women. “Because of contraindications to the use of epinephrine, 10 patients did not receive” 15 [micro sign]g epinephrine. In 13 patients, a concurrent uterine contraction “invalidated” the results of the test dose. Three patients received a second injection of epinephrine because of a “questionable tachycardic response.” Among the 186 patients who received an epinephrine test dose, there were no false‐negative results (upper limit 95% CI, 1.6%), 14 true‐positive results (7.5% of tested catheters, 6.6% of all catheters), and 8 false‐positive tests (4.3%). Thus, although the epinephrine test dose seems unlikely to miss patients with an intravenous catheter, false‐positive results may cause patients to undergo a second epidural insertion unnecessarily.
The Doppler test also produces both false‐positive and false‐negative results. In a study of 303 laboring women, this test detected all 16 intravascular single‐orifice catheters (5.3%). There were no false‐negative tests results (upper limit 95% CI, 1.1%). But six test results were falsely positive (2%). 
A recent study evaluated the Doppler test and multiple‐orifice catheters and it failed to detect two intravascular catheters (upper limit of 95% CI, 1.6%). 
The conclusion that aspiration is too unreliable to be used alone to detect intravascular catheters in laboring women 
is based on studies using epidural catheters with a single, terminal hole. In a survey of 4,003 labor blocks, aspiration was initially negative in 65 of 194 intravascular catheters. 
In another study, initial aspiration detected only 5 of 21 intravenous catheters. 
In contrast, 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…” The data presented here are consistent with this claim. With the possible exception of two atypical catheters, aspiration detected all catheters that were initially inserted into a blood vessel.
We saw two atypical catheters in this study. The first migrated into a blood vessel after initially being placed in the epidural space. Although this catheter may have been in a vessel from the beginning, the clinical presentation suggests otherwise. This patient had some analgesia and bilateral sensory change (including sacral block) 3 h after intrathetical placement of sufentanil. It is unlikely that intrathecal sufentanil alone, or in combination with intravenous bupivacaine‐sufentanil, produced these effects. 
In the second patient, 10 [micro sign]g epinephrine produced maternal tachycardia despite a negative result of aspiration. This result could be a false‐negative aspiration of a false‐positive epinephrine response. Two previous catheters perforated blood vessels. Bleeding and vascular uptake from these sites may have increased the chance of a false‐positive epinephrine response.
The exact classification of these two patients has little effect on the conclusions that can be drawn from this study. The incidence of false‐negative aspiration (0 ‐ 2 in 1,085; upper limit 95% CI, 0.2% to 0.4%) in laboring women is comparable to the reported incidence of false‐negative Doppler tests (0 of 303; upper limit 95% CI, 1.1%) 
and false‐negative epinephrine test doses (upper limit 95% CI, 1.6%). 
False‐positive results occur occasionally with the Doppler (2%) 
and epinephrine (4.3%) tests. 
Unlike other commonly used tests to check for intravenous catheter placement, it is difficult to conceive of a falsely positive aspiration. Clinically, it is important to note that false‐negative results can occur with any test dose. Based on the available data, the frequency of these false‐negative tests is similar when using epinephrine, air, or aspiration alone (Table 5
After blood was noted on initial insertion, six catheters were withdrawn until aspiration was negative. Four of these were positive epidural catheters, but the other two were intravascular on later aspiration. Thus, although aspiration is reliable for detecting initial insertion of a multiple‐orifice catheter into a blood vessel, it is unreliable when used when trying to withdraw a catheter from a blood vessel. In the event of a positive aspiration, we recommend removing and replacing the catheter. If the anesthesiologist tries to withdraw partially and then use the catheter, an additional test dose may be prudent.
It is unclear why a difference exists between single and multiple‐orifice catheters. Tissue may more readily obstruct the hole of an open‐ended, single‐orifice catheter. The blunt end of the multiple‐orifice catheter may serve as a stent within the blood vessel. The presence of multiple holes may increase the chance that one will remain patent. Although these explanations are plausible, they do not account for our observation that aspiration seems more sensitive during initial insertion than when an intravascularly placed catheter is repositioned. Perhaps blunt‐tipped catheters compress the wall of a vein before perforating it. This compression may create a pool of blood that readily flows through the catheter when it enters the vessel. Withdrawing the catheter may then decompress the vein, causing it to collapse and obstruct the lateral catheter holes.
Our results have interesting implications for clinicians who use other test doses in conjunction with aspiration of a multiple‐orifice epidural catheter. Because aspiration alone detects nearly all intravascular catheters, subsequent injection of a test dose is unlikely to yield a true‐positive result. Most other methods of detecting intravascular catheters sometimes produce falsely positive results. In this setting, any positive result is more likely to be a false than a true positive. For example, in a series of 1,000 labor epidural analgesics using a multiple‐orifice catheter, aspiration alone will detect 60 of 62 intravascular catheters (based on the results of this study). Subsequent injection of 15 [micro sign]g epinephrine in the remaining 940 women may detect the two intravenous catheters that did not yield blood on aspiration. However, there also will be 40 falsely positive epinephrine test results. 
Thus, in this setting, the epinephrine test dose, at best, has a positive predictive value (true positives/(true positives + false positives) of only 4.8%.
During this study (October 29, 1996 to June 17, 1997), 1,624 women requested neuraxial analgesia, but only 1,029 were included in the study. However, we believe that the potential for bias played little role in these results. Participation in this protocol was voluntary. Historically, an epinephrine test dose was routinely used at Barnes‐Jewish Hospital. Some anesthetists (M.D. or C.R.N.A.) were uncomfortable omitting this step and declined to participate in this study. Patients not enrolled in the study received the same intrathecal and epidural drugs as those mentioned here. In addition, patients who were not enrolled received an epinephrine‐containing test dose before epidural injection or infusion. This unwillingness to participate in the study also accounts for the greater use of the CSE technique in enrolled patients. Anesthetists who insisted on using an epinephrine test dose also were less likely to incorporate this “new” technique into their practice.
Intrathecal sufentanil and sufentanil + bupivacaine can produce both sensory changes and analgesia. Sensory changes from sufentanil alone regress in <1 h. 
Analgesia usually lasts 60 ‐ 120 min. [22–24]
In patients in advanced labor, sufentanil + bupivacaine will provide pain relief for about 120 min. 
We felt that bilateral sensory change and analgesia present more than 2 h after intrathecal sufentanil injection resulted from the epidural infusion rather than the intrathecal sufentanil.
This study allows us to make some observations about the reliability of catheters inserted as part of the CSE technique. The only nonintravascular catheters that failed completely were inserted as part of an epidural anesthetic. After eliminating intravascular catheters, 93% of catheters (693 of 746) inserted with the CSE technique proved to lie within the epidural space (they were positive epidural catheters). Most of the remaining 53 equivocal epidural catheters could not be reliably assessed because the patients delivered within 2 h of their intrathecal sufentanil.
This, the largest study to date to examine a method to detect intravascular catheters, has shown that aspiration alone routinely detects intravascular multiple‐orifice catheters. These data do not suggest that false‐negative aspiration is impossible, but rather that it is no more likely than a false‐negative response to epinephrine or air. Clinically, these results suggest that, when using a multiple‐orifice catheter, combining observation and aspiration with incremental dosing of local anesthetics is a reasonable method to provide labor epidural analgesia. Under no circumstances, however, should the results of any test of catheter location lead to bolus injection of large, potentially toxic doses of local anesthetics.
1. Kenepp NB, Gutsche BB: Inadvertent intravascular injections during lumbar epidural analgesia. Anesthesiology 1981; 54:172
2. Leighton BL, Norris MC, DeSimone CA, Rosko T, Gross JB: The air test as a clinically useful indicator of intravenously placed epidural catheters. Anesthesiology 1990; 73:610-3
3. Michael S, Richmond MN, Birks RJS: A comparison between open-end (single hole) and closed-end (three lateral holes) epidural catheters. Anaesthesia 1989; 44:578-80
4. Reynolds F: Epidural catheter migration during labour [Letter]. Anaesthesia 1988; 43:69
5. U.S. Government. Part 46-Protection of human subjects. Code of Federal Regulations 1991; title 45:46.116
6. Plumer MH, Rottman R: How anesthesiologists practice obstetric anesthesia. Responses of practicing obstetric anesthesiologists at the 1993 meeting of the Society for Obstetric Anesthesia and Perinatology. Reg Anesth 1996; 21:49-60
7. Beilin Y, Bodian CA, Haddad EM, Leibowitz AB: Practice patterns of anesthesiologists regarding situations in obstetric anesthesia where clinical management is controversial. Anesth Analg 1996; 83:735-41
8. Yoshii WY, Miller M, Rottman RL, Kotelko DM, Wright WC, Stone JJ, Rasmus KT, Rosen PJ: Fentanyl for epidural intravascular test dose in obstetrics. Reg Anesth 1993; 18:296-9
9. Morris GF, Gore-Hickman W, Lang SA, Yip RW: Can parturients distinguish between intravenous and epidural fentanyl? Can J Anaesth 1994; 41:667-72
10. Michels AMJ, Lyons G, Hopkins PM: Lignocaine test dose to detect intravenous injection. Anaesthesia 1995; 50:211-3
11. Colonna-Romano P, Lingaraju N, Braitman LE: Epidural test dose: Lidocaine 100 mg, not chloroprocaine, is a symptomatic marker of iv injection in labouring parturients. Can J Anaesth 1993; 40:714-7
12. Hood DD, Dewan DM, James FM III: Maternal and fetal effects of epinephrine in gravid ewes. Anesthesiology 1986; 64:610-3
13. Leighton BL, Norris MC, Sosis M, Epstein R, Chayen B, Larijani GE: Limitations of epinephrine as a marker of intravascular injection in laboring women. Anesthesiology 1987; 66:688-91
14. Cartwright PD, McCarroll SM, Antzaka C: Maternal heart rate changes with a plain epidural test dose. Anesthesiology 1986; 65:226-8
15. Colonna-Romano P, Salva R, Lingaraju N, Seitman DT: Epinephrine-induced tachycardia is different from contraction-associated tachycardia in laboring patients. Anesth Analg 1996; 82:294-6
16. Colonna-Romano P, Lingaraju N, Godfrey SD, Braitman LE: Epidural test dose and intravascular injection in obstetrics: Sensitivity, specificity, and lowest effective dose. Anesth Anal 1992; 75:372-6
17. Albright GA: Epinephrine should be used with the therapeutic dose of bupivacaine in obstetrics. Anesthesiology 1984; 61:217-8
18. Colonna-Romano P, Lingaraju N: Epinephrine test dose is a reliable test in laboring patients receiving epidural anesthesia: A prospective clinical evaluation study [Abstract]. Anesthesiology 1996; 85:A910
19. Topkis W, Leighton BL, Arkoosh VA, Lee S-H, Huffnagle HJ, Huffnagle S: Effectiveness of the Doppler air test in diagnosing intravascular multi-orifice epidural catheters [Abstract]. Anesthesiology 1997; 87:A883
20. 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
21. Camann WR, Denny RA, Holby ED, Datta S: A comparison of intrathecal, epidural, and intravenous sufentanil for labor analgesia. Anesthesiology 1992; 77:884-7
22. Riley ET, Ratner EF, Cohen SE: Intrathecal sufentanil for labor analgesia: Do sensory change predict better analgesia and greater hypotension? Anesth Analg 1997; 84:346-51
23. Norris MC, Grieco WM, Borkowski M, Leighton BL, Arkoosh VA, Huffnagle HJ, Huffnagle S: Complications of labor analgesia: Epidural versus combined spinal epidural techniques. Anesth Analg 1994; 79:529-37
24. Campbell DC, Camann WR, Datta S: The addition of bupivacaine to intrathecal sufentanil for labor analgesia. Anesth Analg 1995; 81:305-9
25. Viscomi CM, Rathmell JP, Pace NL: Duration of intrathecal labor analgesia: Early versus advanced labor. Anesth Analg 1997; 84:1108-12
© 1998 American Society of Anesthesiologists, Inc.