Secondary Logo

Journal Logo

Original Article

Lumbar epidural catheter insertion: the midline vs. the paramedian approach

Leeda, M.1; Stienstra, R.1; Arbous, M. S.1; Dahan, A.1; Th. Veering, B.1; Burm, A. G. L.1; Van Kleef, J. W.1

Author Information
European Journal of Anaesthesiology: November 2005 - Volume 22 - Issue 11 - p 839-842
doi: 10.1017/S0265021505001419



Continuous epidural anaesthesia (CEA) is a technique frequently used to provide anaesthesia for operative procedures. The epidural catheter offers the possibility to prolong anaesthesia during surgery when necessary and in the postoperative period the catheter may be used to provide analgesia. Although CEA is a well-established technique, the insertion of an epidural catheter is not without complications or side effects.

The most frequent side effect related to epidural catheter insertion is the occurrence of paraesthesias, with a reported incidence varying from 16 to 89% [1-4]. Although paraesthesias during epidural catheter insertion are usually mild and not associated with untoward sequelae, they may also signal nerve root damage. Moreover, they may be experienced as unpleasant by the patient. Lowering the incidence of paraesthesias would therefore be beneficial.

The lumbar epidural space may be entered via a midline or a paramedian approach. The midline approach is used most widely, probably because it requires less three-dimensional insight and the ligamentum flavum is at its widest in the midline and therefore easily identified in the majority of patients. However, the paramedian approach may have advantages over the midline approach: it is our experience that in case of a narrow lumbar interspace, the paramedian route offers a higher success rate as it bypasses most of the bony structures impeding the advancement of the epidural needle in the midline. It is conceivable that the steeper angle in the paramedian approach facilitates epidural catheter insertion. It has been suggested that the incidence of accidental dural puncture is lower when using the paramedian route [5] and in one study comparing the midline and the paramedian approach, the latter was associated with fewer technical problems [6].

The purpose of this study was to compare the characteristics of epidural catheter insertion via the midline or the paramedian approach with regard to ease of catheter insertion, incidence of paraesthesias and efficacy of sensory and motor blockade. In addition to the type of approach, the prognostic value of Patients characteristics variables with regard to the incidence of paraesthesias was assessed.


Thirty patients (aged 18-80 yr, ASA I-III; non-pregnant) scheduled for lower limb surgery under lumbar epidural anaesthesia were studied. The study protocol was approved by the local Ethics Committee of the Leiden University Medical Centre and informed consent was obtained from all patients. Using a computer-generated randomization list, the patients were allocated to either the midline or the paramedian group, each group containing 15 patients.

Premedication consisted of midazolam 7.5-15 mg orally (dose based on patient's age and physical condition) 1 h before regional anaesthesia was scheduled. After intravenous (i.v.) access was established, standard routine monitoring (continuous electrocardiogram, non-invasive blood pressure measurements at 5 min intervals and peripheral oxygen saturation) was applied throughout the anaesthetic and surgical intervention. An additional i.v. dose of 1-2 mg midazolam was administered immediately prior to lumbar puncture. Epidural anaesthesia was performed with the patient in the sitting position. All lumbar punctures were performed by one investigator (RS) experienced with both the midline and the paramedian approach to the epidural space. After local infiltration with lidocaine, a 17 G Tuohy needle (Becton Dickinson, Franklin Lakes, NJ) was introduced through the skin at the third lumbar interspace. The stylet was removed and a 10 mL syringe filled with normal saline was attached to the epidural needle. At that time, the investigator performing the lumbar puncture signalled his readiness to a blinded observer (positioned in front of the patient and unable to see the procedure) by saying the word ‘Start’. The blinded observer responded by starting a stopwatch and confirm the word ‘Start’, at which time the investigator performing the lumbar puncture would start advancing the epidural needle. Identification of the epidural space by loss of resistance was signalled by saying ‘Stop’ at which time the blinded observer stopped the stopwatch and noted the time. Care was taken to limit the volume of saline injected on entry into the epidural space to <1 mL. After identification of the epidural space and a negative aspiration test, a 19 G multi-orifice catheter was inserted 5 cm into the epidural space and the epidural needle was removed. The time needed to introduce the epidural catheter was measured in a similar fashion by saying ‘Start’ when the tip of the catheter was introduced into the hub of the epidural needle and ‘Stop’ when the mark corresponding to 5 cm reached the hub of the epidural needle. During catheter insertion the patient was carefully observed for signs of paraesthesia. Immediately after epidural catheter insertion, the blinded observer asked the patient for any signs of pain or discomfort. After a negative aspiration test and a negative test dose of 3 mL prilocaine 1% with epinephrine an epidural loading dose of 15 mL ropivacaine 1% was administered. The patient was then returned to the supine horizontal position and the time was designated as t = 0.

The upper level of sensory blockade was measured every 5 min by the blinded observer using loss of temperature sensation in the anterior axillary line with an ice cube. The highest segment at which the patient was not capable of detecting the cold temperature of the ice cube was recorded. Establishment of the maximum level of sensory blockade was defined as no further increase during three consecutive measurements and more than 20 min after epidural injection. Motor blockade of the lower limbs was scored according to a 12-point scale, where each joint of the lower limbs (hip, knee, ankle) was scored from 0 to 2 (0: no motor block; 1: partial motor block; 2: complete motor block). The maximum motor block score was evaluated at 5 min intervals until maximum motor block had been established or until 30 min after epidural injection. Quality of surgical anaesthesia was judged by the need for analgesics during surgery and by the patient after surgery on a 3-point scale (good/fair/poor).

The following outcome variables were studied:

The time needed to identify the epidural space;

The time needed to site the epidural catheter as a measure for ease of catheter insertion;

The incidence of paraesthesias;

Block characteristics;

The need for intraoperative analgesics;

The verbal rating of the quality of analgesia after surgery.

Statistical analysis

Patient characteristics and continuous variables were compared using the U-test, the distribution of gender and ASA classification between the two groups were compared using χ2 analysis with Yates correction. To derive a prognostic model for the occurrence of paraesthesia, type of approach and patient characteristics variables (age, height, weight, gender) were considered potential prognostic variables. The association between the incidence of paraesthesias and the potential prognostic variables was explored using univariate and multivariate logistic regression. Variables showing P <0.15 in the univariate analysis were considered potential prognostic varables for the incidence of paraesthesias according to the methods described by Hosmer and colleagues [7]. These were entered, with type of approach, in a multivariate prognostic model. Interaction of biologically plausible combinations of the intervention and other determinants, or of determinants with each other were also tested. A P-value <0.05 was considered significant.


There were no significant differences among the two groups regarding age, height or weight (Table 1). Although not statistically significant, the gender distribution was uneven. The time needed to identify the epidural space was 13.2 ± 5.4 s for the midline and 21.1 ± 19.7 s for the paramedian approach. Although statistically not significant, the apparent difference in the time needed to identify the epidural space was caused by one patient in the paramedian group in whom identification of the epidural space was difficult and took 65 s, due to several redirections of the epidural needle. Epidural catheter insertion was significantly faster in the paramedian group (9.0 ± 5.1 s vs. 18.2 ± 6.2 s, P <0.001). Five patients in the midline group experienced paraesthesias during catheter insertion as opposed to one patient in the paramedian group. In the univariate analysis, this difference was n.s. However, in the multivariate analysis, type of approach (midline vs. paramedian) was a significant independent predictor of paraesthesias. A non-significant trend was found towards a higher risk in female patients (Table 2).

Table 1
Table 1:
Patient characteristics.
Table 2
Table 2:
Multivariate analysis.

The epidural test dose was negative in all patients. Epidural block characteristics in terms of maximum level of sensory blockade and maximum motor block score were similar. Median maximum levels of sensory blockade reached T4 (range C3-L1) in the midline group and T4 (range T2-T12) in the paramedian group. Median maximum motor block scores were 4 (range 0-10) and 6 (range 0-12) in the midline and paramedian groups respectively. The quality of sensory and motor blockade was adequate in both groups. Surgery was completed under epidural anaesthesia in all patients, with none of the patients requiring general anaesthesia or i.v. opioid supplementation. All patients in the midline group had a good quality of block whereas one in the paramedian group reported a fair quality of block. In no case was the block poor.


We found a higher incidence of paraesthesias in the midline group (33% vs. 6.7%). Although this difference was n.s. in the univariate analysis, the multivariate analysis demonstrated that type of approach was a significant independent predictor of paraesthesia. A trend was found towards female gender as a risk factor for paraesthesias during epidural catheter insertion.

A higher incidence of paraesthesias when using the approach has been reported by others. In a retrospective study, Jaucot [8] found a lower incidence of paraesthesias (22% vs. 42.5%) and of vascular puncture when the paramedian approach was used. Blomberg and colleagues [6] reported an incidence of 36% with the midline approach as opposed to 4% with the paramedian approach. Our findings are very similar to those of the latter study. Jaucot [8] found a relatively high incidence of paraesthesias associated with the paramedian approach (22%) as compared to the incidence found by Blomberg and colleagues [6] and by us. Also, Jaucot found paraesthesias to be severe in 5.6% of the midline catheter insertions and 1.5% of the paramedian catheter insertions, whereas in our population no severe paraesthesias were observed. The reason for these differences remains speculative; Jaucot's study was carried out between 1981 and 1985 and differences in catheter stiffness may play a significant role. We have no explanation for our finding that gender may be a prognostic variable with regard to the occurrence of paraesthesias. The reduced incidence of paraesthesias associated with paramedian catheter insertion may be explained by our finding that insertion is easier when using the paramedian route.

Our results show that catheter insertion was significantly faster and thus easier when using the paramedian approach. A similar finding was reported by Blomberg and colleagues [6], who found resistance to catheter insertion less common with the paramedian as compared to the midline approach. The most likely explanation for this observation is the steeper angle of entry of the paramedian epidural needle into the epidural space, facilitating catheter insertion. In an epiduroscopic cadaver study [5], it was demonstrated that when the catheter is advanced in the midline, there is considerable dural tenting and the course of the catheter is unpredictable due to strands of connective tissue restricting movement of the dura mater; on the other hand, paramedian catheters were advanced easily in a cephalad direction without dural tenting.

The advantages of the paramedian approach are not undisputed. Although the paramedian route may be more successful in patients with a narrow lumbar interspace, the required three-dimensional insight may be more difficult for the inexperienced anaesthesiologist. In a study assessing the epidural success rate of new residents, the midline approach was associated with fewer attempts and a higher success rate [9]. On the other hand, Sprung and colleagues [10] found no differences in success rate between the paramedian and midline approach. The presumably lower risk of accidental dural puncture when using the paramedian approach is based on a cadaver study [5] and has not been substantiated by clinical studies to date.

We found no differences in the quality and characteristics of sensory and motor blockade. All patients termed the quality of analgesia as good with the exception of one patient in the paramedian group who scored the quality of analgesia as fair. The reason for this was that he was aware of the surgical manipulation of his ankle joint, although this was not painful and required no additional analgesia. Interestingly, motor block was not complete in the majority of patients, despite our use of the 1% solution of ropivacaine. However, it should be emphasized that our observation period of sensory and motor block was limited to 30 min after epidural injection and motor block may have intensified beyond this point in time.

In summary, under the conditions of our study epidural catheter insertion was significantly faster when using the paramedian approach. In the multivariate analysis, the type of approach was an independent significant predictor of the occurrence of paraesthesias. There was a trend towards a higher incidence of paraesthesias in female patients, especially when using the midline approach.


Supported by the Department of Anesthesiology, Leiden University Medical Center, Leiden, The Netherlands.


1. Rolbin SH, Hew E, Ogilvie G. A comparison of two types of epidural catheters. Can J Anaesth 1987; 34: 459-461.
2. Scott DA, Beilby DS. Epidural catheter insertion: the effect of saline prior to threading in non-obstetric patients. Anaesth Intens Care 1993; 21: 284-287.
3. Hetherington R, Stevens RA, White JL, et al. Subjective experiences of anesthesiologists undergoing epidural anesthesia. Reg Anesth 1994; 19: 284-288.
4. Munoz HR, Dagnino JA, Allende M et al. Direction of catheter insertion and incidence of paresthesias and failure rate in continuous epidural anesthesia: a comparison of cephalad and caudad catheter insertion. Reg Anesth 1993; 18: 331-334.
5. Blomberg RG. Technical advantages of the paramedian approach for lumbar epidural puncture and catheter introduction. A study using epiduroscopy in autopsy subjects. Anaesthesia 1988; 43: 837-843.
6. Blomberg RG, Jaanivald A, Walther S. Advantages of the paramedian approach for lumbar epidural analgesia with catheter technique. A clinical comparison between midline and paramedian approaches. Anaesthesia 1989; 44: 742-746.
7. Hosmer DW. Model building strategies. In: HosmerDW, LemeshowS, eds. Applied Logistic Regression, 2nd edn. New York, USA: John Wiley & Sons, 2000: 91-142.
8. Jaucot J. Paramedian approach of the peridural space in obstetrics. Acta Anaesthesiol Belg 1986; 37: 187-192.
9. Kopacz DJ, Neal JM, Pollock JE. The regional anesthesia ‘learning curve’: What is the minimum number of epidural and spinal blocks to reach consistency? Reg Anesth Pain Med 1996; 21: 182-190.
10. Sprung J, Bourke DL, Grass J et al. Predicting the difficult neuraxial block: A prospective study. Anesth Analg 1999; 89: 384-389.


© 2005 European Society of Anaesthesiology