Success of Spinal and Epidural Labor Analgesia: Comparison of Loss of Resistance Technique Using Air versusSaline in Combined Spinal-Epidural Labor Analgesia Technique

After securing Institutional Review Boards (Forsyth Medical Center Institute Review Board, Winston-Salem, NC, and Wake Forest University Institute Review Board, Winston-Salem, NC) approval and written informed consent, 360 women in active labor and requesting neuraxial labor analgesia were randomized to have CSE placement using either air or saline in LORT. Inclusion criteria were: cervical dilation of no more than 8 cm, verbal rating score for pain of at least 6 of 10 maximum, vertex singleton pregnancy, and no medical/obstetric contraindications for CSE placement. Exclusion criteria were: American Society of Anesthesiologists physical status greater than II or weight greater than 114 kg. Computer-generated randomization was used for randomized group assignment. The randomized group assignment was placed sequentially on pages inside a notebook, from which the anesthesiologist performing the procedure obtained the randomized group assignment for the current subject sequentially and without the knowledge of the previous or future subject group assignment. Except for the anesthesiologist performing the CSE procedure, the patient, the nurses, the data collector, and interpreters were blinded to the group assignment. Upper level anesthesia residents with over 2 months of successful obstetric anesthesia training experiences, obstetric anesthesia fellows, and attending obstetric anesthesiologists, all of whom are comfortable with the CSE procedure using either air or saline LORT, participated in placing the blocks. All trainees were supervised by an attending obstetric anesthesiologist.

The patient was placed in the sitting position. After the back was sterilely prepared and draped, 1% lidocaine was injected subcutaneously for local anesthesia in the lower lumbar vertebral interspace between L2 and L5 and chosen by the anesthesiologist performing the procedure after examining patient's back. The epidural space was located using 3 ml of air (group Air) or 3 ml of saline (group Saline) for the LORT with a glass syringe and a 17-gauge, 89-mm Weiss epidural needle (Becton Dickinson and Company, Franklin Lakes, NJ). With identification of the epidural space, the full 3-ml volume in the syringe was injected into the epidural space according to their randomized group assignment. If part of the volume in the syringe was lost during or before successful identification of epidural space, the syringe was refilled back to 3 ml with air or saline for injection according to the group assignment before successful identification of the epidural space. The 3-ml volume used and injected per our study design was to ensure consistency and control in volume injected epidurally between patients and groups. Using the needle through needle technique, a noninterlocking 27-gauge, 119-mm Whitacre spinal needle (Becton Dickinson and Company) was advanced through the Weiss needle until the hubs touched to reach the subarachnoid space. The distal tip of the spinal needle extended 13 mm past the distal tip of the Weiss epidural needle. Spontaneous fluid return to the needle hub as well as presence or absence of fluid return from preinjection and postinjection aspiration for CSF using the 3-ml syringe used for administering the spinal dose was recorded. The presence of spontaneous fluid returning to the spinal needle hub was considered as identification of the intrathecal space. Once identified, the spinal dose administered consisted of 0.6 ml of bupivacaine 2.5 mg/ml with 0.4 ml of fentanyl 50 μg/ml. After removal of the spinal needle, a 19-gauge, triport epidural catheter (Becton Dickinson and Company) was then inserted and secured with 5–6 cm left inside the epidural space. The patient was placed in a position with a folded up blanket (roll) placed behind the hip and lumbar areas to keep tilted laterally to the left or right position as she felt comfortable in and as tolerated by both the patient and fetus. No epidural medications were administered until 15 min after the spinal dose administration.

Fetal heart rate, maternal hemodynamics, verbal pain score (0 indicating no pain and 10 indicating the worst pain imaginable) and verbal pruritus score (0 indicating no pruritus and 10 indicating worst pruritus imaginable) were recorded before the CSE procedure and then every 5 min for 20 min after spinal needle removal by the anesthesiologist performing the CSE procedure. Another anesthesiologist or research nurse blinded to the group assignment collected the data for the rest of the study after the initial 20 min of the CSE procedure. Hypotension (defined as systolic blood pressure decrease greater than 20% from preprocedure blood pressure) was treated with increments of 5 to 10 mg of ephedrine per usual practice. Evaluation of the spinal dose efficacy consisted of verbal pain score and sensory level to cold temperature.

If blood returned spontaneously or with aspiration through the epidural catheter at any time, the catheter was injected with 0.5 ml of saline incrementally and pulled back 0.5 cm incrementally if needed until the catheter was cleared of blood or with 3 cm of catheter remained in epidural space. If unsuccessful, the catheter was removed, and a new epidural catheter was inserted using the same LORT (3 ml of air or saline) as previously assigned. The absence or presence of paresthesia at the time of insertion of the spinal needle or epidural catheter was documented. Paresthesia was defined as a tingling or mild electrical shock sensation or the sensation of hitting a funny bone but of less or same intensity. If paresthesia did not resolve and lasted more than 1 min, procedure was abandoned and patient was excluded from the study.

After the initial 15 min, an intrathecal test dose of 2 ml of lidocaine 20 mg/ml was administered via  the epidural catheter; if negative, epidural patient-controlled analgesia was initiated in the usual manner by using bupivacaine 1.25 mg/ml and fentanyl 2 μg/ml with a basal infusion rate of 12 ml/h, a demand dose of 6 ml, and a 10-min lockout. Patients were assessed for pain level every 2 h after initiation of patient-controlled epidural analgesia or more frequently if patient reported a pain score of greater than 3 or required physician-administered boluses until verbal pain score reduced to no more than 3. For inadequate labor analgesia (defined as verbal pain score greater than 3), catheter adjustment by pulling catheter back to 3 cm inside the epidural space if unilateral block noted or additional physician administered boluses of bupivacaine 2.5 mg/ml first, and then lidocaine 20 mg/ml if needed, both given in increments of 5 ml up to a total of 10 ml each, were allowed to achieve adequate analgesia as needed. If analgesia was still inadequate (verbal pain score greater than 3) after physician-administered boluses and catheter manipulation, epidural catheters were replaced, and the time and reason of replacement were documented. The total number of bolus doses was determined, as was the presence or absence of sensory levels and verbal pain score during the physician-administered boluses up to a maximum of 20 ml.

In the case in which there was no spontaneous clear fluid return into the spinal needle hub after insertion of the spinal needle and removal of spinal needle stylet, spinal dose was not administered, and a triport epidural catheter was inserted 5–6 cm inside the epidural space after spinal needle removal as in usual manner. The lack of spontaneous clear fluid return was defined as failed spinal analgesia by our study design. Then, an intrathecal test dose of 2 ml of lidocaine 20 mg/ml was administered followed 5 min later by an intravenous test dose of 5 ml of lidocaine 20 mg/ml if the intrathecal test was negative. The rest of the epidural patient-controlled analgesia and breakthrough pain management was the same as described above.

The primary outcome measure was the success of spinal labor analgesia as defined by verbal pain score of no more than 3 at 15 min after spinal dose administration. Secondary outcomes were spontaneous fluid return, fluid return upon preinjection and postinjection aspiration, epidural catheter replacement rate, and the average hourly total amount of epidural medications required within 4 h from initial epidural catheter insertion. The average hourly total amount (ml/h) of epidural drug used (reported as ml/h in equivalents of bupivacaine 1.25 mg/ml with fentanyl 2 μg/ml) was calculated as the sum total of the epidural patient-controlled analgesia basal infusion amount, demand doses, and physician-administered boluses divided by the actual duration of consumption up to 4 h. For the purpose of this calculation, we estimated the local anesthetic-sparing effect of fentanyl as follows: (1) bupivacaine 1.25 mg/ml estimated as 0.85 equivalents of bupivacaine 1.25 mg/ml with fentanyl 2 μg/ml; (2) bupivacaine 2.5 mg/ml estimated as 1.70 equivalents of bupivacaine 1.25 mg/ml with fentanyl 2 μg/ml; (3) lidocaine 20 mg/ml estimated as similar to bupivacaine 5 mg/ml and 3.40 equivalents of bupivacaine 1.25 mg/ml with fentanyl 2 μg/ml. Epidural catheter efficacy was measured by requirement (or its lack of) for epidural catheter replacement, physician-administered supplement boluses, and the average hourly total amount of epidural drug used within 4 h of analgesia initiation.

On the basis of our quality assurance data on CSE success rate and previous published reports on success rate of CSE spinal labor analgesia ranging from 83 to 99%, we considered a difference between 98% versus  90% in success rate between groups to be clinically significant and relevant.7,8 A priori  power analysis revealed that a sample size of 160 per group was needed to demonstrate the stated difference (90% vs . 98%) between groups in the primary outcome variable to achieve a power of 0.8 and an alpha of 0.05. An estimated total of 360 subjects was planned for enrollment, with the goal to obtain at least 320 evaluable subjects. A separate subgroup analysis was also planned a priori  to compare spinal analgesia success and/or epidural catheter replacement rate between patients with different CSE characteristics (those with initial clear fluid returned to spinal needle vs . those without and those with initial clear fluid return and with positive aspiration before and/or after spinal drug administration vs . those with negative aspiration).

For statistical analysis, SigmaStat version 3.01 for Windows version (SPSS Inc., Chicago, IL) was used. Incidence and proportion data were compared between groups using chi-square test or Fisher exact test. The average hourly total epidural drug consumption was compared between groups using 2-tailed unpaired t  test. For continuous data that did not meet parametric assumptions, the Mann-Whitney U  test was used. Change in continuous measures over time within each group was determined using one-way analysis of variance for repeated measures and two-way analysis of variance for comparison between groups with planned post hoc  Tukey test as needed. P < 0.05 is considered statistically significant. Data are presented as mean values ± SD, median values, interquartile range, and percentages as appropriate.

Of the 360 laboring patients enrolled, 15 were excluded, resulting in the remaining 173 patients in the Air group and 172 in the Saline group being included for statistical analysis. Among those 15 excluded patients, 3 had unintentional dural puncture (2 in Air group, 1 in Saline group), 1 had intrathecal catheter (in Air group), 3 delivered rapidly within 15–20 min after administration of spinal drug (all in Air group), 1 did not receive spinal drug injection (Saline group) because of significant paresthesia during spinal needle insertion, 1 had a protocol violation (due to premature administration of epidural test doses), and 2 had incomplete data set (both in Saline group), and 4 were consented but not studied. None of the patients with unintentional dural puncture had resulting clinical pneumocephalus.

Patient demographics, obstetric data, and anesthesia provider experience were comparable in the two treatment groups (table 1). The overall rate of successful labor spinal analgesia for all subjects was 93.9%, and there was no significant difference between Air (94.2%) and Saline (93.6%) groups. There were seven patients who had no initial spontaneous clear fluid return (4 in Saline group and 3 in Air group) through the spinal needle, and the epidural catheter was inserted without having spinal drug administered as per our study protocol. None of these three patients in the Air group required epidural catheter replacement, whereas two of the four patients in the Saline group required epidural catheters replacement subsequently even though they provided adequate analgesia at the initiation of the epidural analgesia. When excluding those seven patients without initial spontaneous clear fluid return from spinal analgesia failure, the overall rate of successful labor spinal analgesia after first observing clear fluid return to spinal needle hub was 95.9% for all subjects, without significant differences between groups (95.9% in Air group vs . 95.8% in Saline group; table 2).

The incidence of epidural catheters requiring replacement within the first 4 h after initial epidural catheter insertion was similar between Air and Saline groups (2.9% or 5 of 173 in Air group vs . 5.2% or 9 of 172 in Saline group, 4.1% or 14 of 345 for all patients combined). Also similar was the amount of the average hourly total epidural medication required for labor analgesia (18 ± 6 ml/h vs . 19 ± 7 ml/h in equivalents of bupivacaine 1.25 mg/ml with fentanyl 2 μg/ml for Air and Saline groups, respectively) and percent of patients requiring physician-administered boluses (19.7% or 34 of 173 in Air vs . 17.4% or 30 of 172 in Saline group).

When including only those patients with initial spontaneous clear fluid return to spinal needle hub, the occurrence of negative continuous fluid aspiration before spinal drug injection (11.8% or 20 of 170 in Air group vs . 9.5% or 16 of 168 in Saline group; overall 10.7% or 36 of 338) or after spinal drug injection (13.5% or 23 of 170 in Air group vs . 11.3% or 19 of 168 in Saline group; overall 12.5% or 42 of 338) were similar between groups.

Among those with positive initial spontaneous fluid return to the spinal needle but negative continuous fluid return on preinjection aspiration (20 in Air group, 16 in Saline group, 36 overall), the overall spinal analgesia success rate (94.4% or 34 of 36) after spinal drug administration among these 36 patients was similar to the overall success rate (95.9%) of all 338 patients (including all patients both with positive and negative fluid return upon aspiration after the initial spontaneous clear fluid returned to the spinal needle hub). Twenty patients in the Air group had negative continuous fluid return on aspiration after initial spontaneous fluid return, but none had failed spinal analgesia or required epidural catheter replacement after spinal drug administration, whereas 2 of 16 patients in the Saline group with same setting had failed spinal analgesia and required subsequent epidural catheter replacement within 4 h after spinal drug administration. However, these differences between groups were not statistically significant.

Similarly, among those with positive initial fluid return to the spinal needle but negative continuous fluid return on postinjection aspiration (23 in Air group, 19 in Saline group, 42 overall), the overall spinal analgesia success rate was 97.6% and was not different from that of all 338 patients. Among those without initial clear fluid return to spinal needle hub, the epidural catheter replacement rate within first 4 h was 28.6% (95% CI 7.6–65%) and was significantly (P < 0.03) higher than the 3.6% (95% CI 1.98–6.17%) rate among those 338 patients with presence of initial clear fluid return to the spinal needle hub.

During the initial epidural catheter placement, blood returned either spontaneously or with aspiration in 11 (3.2%) of the epidural catheters (6 in Air group, 5 in Saline group), among which 9 of them (5 in Air group and 4 in Saline group) were cleared of blood with maneuvers described in the Methods section and provided adequate analgesia without the need for subsequent replacement within the study period. The remaining two catheters were replaced, and new catheters were reinserted using the same LORT technique (3 ml of air or saline) as previously assigned, and both provided adequate analgesia for the duration of the study period. Among those few patients without initial spontaneous fluid return, there was no inadvertent intravenous catheter. There was no excessively high block or total spinal among our subjects. All paresthesia occurred with the spinal needle or epidural catheter insertion was transient except in one excluded subject. The incidences of paresthesia with the spinal needle insertion were 6.2% for Air group and 6.3% for Saline group and 24% for both groups with epidural catheter insertion, all without differences between groups.

The results of this study demonstrate three clinically relevant findings that have not been previously studied in a prospective controlled manner but related to clinical situations and decisions encountered almost daily by obstetric anesthesiologists. The first is that the success of spinal analgesia, the success of obtaining initial spontaneous fluid return through the spinal needle, and subsequent epidural catheter efficacy with needle through needle CSE technique is the same regardless of using saline or air for LORT in identifying the epidural space. Second, the likelihood of successful spinal analgesia in the setting when clear fluid cannot be aspirated before or after spinal drug injection after positive initial spontaneous fluid return to the spinal needle hub was the same as those with fluid return upon preinjection and/or postinjection aspiration. Third, the risk of epidural catheter failure is significantly higher if no initial clear fluid returned to the spinal needle during the CSE technique.

Lack of spontaneous CSF flow at the spinal needle hub on initial placement accounted for 1.7% in the Air group and 2.2% in the Saline group, which are similar to the previously reported failure rates of 0.5–5%.10–12The much lower incidence of failed CSF flow before spinal drug administration reported by Albright et al . could be explained in part by the retrospective nature of their report and their practice of redirecting the spinal needle if CSF was not obtained upon first pass, which we did not do by our study design.12Of note, Albright et al . injected sufentanil up to 20 μg with or without bupivacaine through the spinal needle in 11 patients in whom they were unable to obtain “continuous” CSF flow. Of the eleven patients, eight reported some pain relief. However, the report did not specify if any initial spontaneous CSF fluid return was observed. Furthermore, even an inadvertent epidural administration of 20 μg of sufentanil together with its associated systemic absorption may provide significant initial pain relief.13In our study, we considered the absence of initial fluid return as spinal analgesia failure and did not administer any intrathecal drug.

The spinal analgesia was also considered to have failed if the patient did not obtain adequate pain relief from the spinal dose. This occurred in 4.1% and 4.2% of patients in the Air and Saline groups, respectively. The rate of spinal anesthesia failure for surgical patients noted to have free CSF flow has been reported between 2.0% to as high as 15%.7,8Our failure rate is in the low range of this, possibly resulting from the sympathetic nature of labor pain, which does not require the same level of block density as surgery. It is interesting to note that among all 14 patients who had failed spinal analgesia with positive initial spontaneous fluid return to the spinal needle hub, continuous fluid return was observed in 12 of the 14 patients at both preinjection and postinjection aspiration. One can only speculate that these failures (especially for the Air group in which no fluid was injected before the spinal needle insertion) may possibly be due to maldistribution, anatomical abnormalities, individual differences in pain and analgesic sensitivity, and varying intensity of labor pain and progression.

Given the very similar success rate of spinal analgesia and epidural efficacy between Air and Saline groups, it is evident that 3 ml of saline injected epidurally did not seem to cause misidentification of spinal fluid return from intrathecal space when spinal needle was inserted during a CSE procedure. Other studies have suggested both risks and benefits in using saline versus  air in LORT for traditional epidural placement.2We used 3 ml of air or saline to standardize the volume injected epidurally in the study for control comparison between groups. In our clinical practice, when air is used, practitioners often inject less than 0.25 to 0.5 ml when change or loss of resistance is identified. This study does not rule out the possibility of misidentification of fluid as CSF if a larger amount of saline was injected epidurally during LORT. Three milliliters of saline may have been too small a volume to cause maldistribution of local anesthetic in the epidural space or tenting of the dura, which has been shown to occur during the CSE technique in cadaver epiduroscopy studies and has been theorized to prevent the spinal needle from reaching the intrathecal space.14,15The acceptable length for spinal needle extension from the tip of the epidural needle has been established as 10–15 mm. Longer needles can lead to increased paresthesias with insertion, and shorter needles would cause failure to obtain spinal fluid.15Our spinal needles extend 13 mm from the tip of the Weiss needle, leading us to believe that this combination should lead to adequate results.

Even when there was no continuous fluid return upon aspiration after observing initial spontaneous fluid return to the spinal needle hub, an overall 94.4% success rate of spinal analgesia performed with a needle through needle CSE was observed. In a search of the literature, this outcome in CSE technique is a clinically significant question that has not been previously studied in a prospective controlled manner. Occurrence of positive initial spontaneous fluid return with subsequent absence of fluid upon aspiration is not uncommon in daily practice, and the chance of success of the subsequent intrathecal dose has previously been unknown. As long as initial clear spontaneous fluid was observed at the spinal needle hub, our data demonstrated a similarly high success rate of spinal analgesia. Therefore, our data suggest the practice of aspiration before injection may be unnecessary because it may not yield any additional useful clinical information to the outcome of the spinal analgesia. We do not know if the success rate would be the same in the similar scenario for spinal anesthesia instead of analgesia under CSE or spinal anesthesia alone.

Another interesting piece of information this study yielded is that there is a relatively high rate of negative postinjection aspiration of 12.4% (42 of 338 all combined, 23 in Air group, 19 in Saline group), even in the presence of initial spontaneous fluid return to spinal needle at spinal needle insertion. All of these 23 patients in Air group and 18 of those 19 patients in Saline group had successful spinal analgesia despite negative postinjection aspiration. Munhall et al . showed that 2 (1%) of 198 spinal anesthetics had negative postinjection aspiration, and both patients had adequate surgical spinal anesthesia.16The lower incidence of negative postinjection aspiration in their study may have been due to larger needle size (18- to 25-guage) and the use of a spinal technique versus  the CSE technique. In a prospective study of 1,891 patients, Tarkkila et al . found the failure rate of spinal anesthesia with CSF flow present postinjection in a surgical setting was 2.9%. The failure rate increased to 3.8% if postinjection aspiration was negative. However, these differences were insignificant, and their overall failure rate was 3.1%.17These findings were similar to ours showing that confirmation of CSF flow during or after injection may also be an unnecessary practice. However, in our study, this group of patients was small, and the significance of these findings would have to be confirmed in an investigation with a larger sample size.

The epidural catheter replacement rate was significantly (P < 0.03) higher among those patients when epidural catheter was inserted after not being able to first obtain spontaneous CSF to the spinal needle hub in the CSE technique. Of these patients, 28.6% required their epidural catheters replaced within 4 h from initial epidural catheter placement, whereas all patients included in the Air and Saline groups had only an overall 2.9% and 5.2% of patients requiring replacement. Among all the patients with the presence of initial fluid returned in the spinal needle, the overall epidural catheter replacement rate was only 3.6%. This result confirms and is comparable to our previous study, suggesting a 22% epidural catheter replacement rate during labor among similar patients in similar settings when no spontaneous initial CSF flow appeared at the spinal needle hub during needle through needle CSE technique.18When data from this previous study are combined with the current study, the incidence of epidural catheter replacement rate among those without initial clear fluid return to the spinal needle during a labor CSE procedure is 24% (6 of 25, 95% CI 11.2–43.7%) and is still significantly higher than the 3.6% (95% CI 1.98– 6.17%) among all patients with initial fluid return (P < 0.0006). The absence of initial spontaneous fluid return after insertion of spinal needle in a CSE procedure may potentially indicate a high risk of failure in the epidural catheter inserted without repeating the epidural procedure. Collectively, this strongly suggests reidentification of the epidural space at a different angle or interspace in such settings is warranted. Despite statistically significant differences observed, both the number of patients in this setting from the current study and our previous study are small, further studies are needed to confirm this finding.

The use of saline versus  air for LORT with traditional epidural technique has been debated for years and been addressed often with nonconclusive reports of air-associated “patchy” blocks, wet taps and even subcutaneous emphysema.2,19–21Our study on CSE showed no significant difference in the incidence of spinal analgesia success, epidural efficacy, and dural puncture between Air and Saline groups. However, we chose only providers accustomed to using both air and saline for the LORT.

Previous studies had evaluated air versus  saline LORT only in traditional epidural techniques but not in CSE or spinal analgesia. Valentine et al . were the first to report a controlled study using 4 ml of air or saline for LORT with traditional epidural technique and epidural analgesia. The air group had a higher incidence of unblocked dermatomal segments that were completely relieved with an additional dose of medication in both groups without the need of catheter replacement.22Similar findings were elicited in another study by Beilin et al . in evaluating the success in initiation of traditional epidural analgesia.23Our study shows that whether 3 ml of air or saline is used for the LORT with CSE did not affect inadvertent dural puncture rate, number of catheter replacements, or how much epidural medication the patients required over the 4-h study period. The decision of using air or saline LORT for CSE labor analgesia should therefore not be based on spinal analgesia success or epidural catheter efficacy; rather, it should be based on other reasons and outcomes with scientific evidences as well as the practitioner's experience level with air or saline LORT.

Our study shows that there is no difference in the failure rate of spinal analgesia or efficacy of the epidural catheter function when using either air or saline in loss of resistance technique for CSE labor analgesia. This study also suggests that successful spinal analgesia was obtained as frequently in the setting of initial spontaneous fluid flow return without the ability to aspirate as in those patients where aspiration was possible. This cannot be extrapolated to spinal anesthesia for surgery and should be further studied in a larger population. We also determined that the risk of epidural catheter failure is significantly higher if spontaneous CSF flow is not obtained upon passing the spinal needle in a CSE technique. However, this was a small subset of patients, and larger studies with this finding as a primary outcome would be needed to confirm this finding.