Access to the PPM Journal and newsletters is FREE for clinicians.
11 Articles in Volume 6, Issue #7
An Overview of Sleep Medications
Editor's Memo
Ernest Syndrome and Insertion of the SML at the Mandible
Low Level Laser Therapy – A Clinician’s View
Microcurrent Electrical Therapy (MET): A Tutorial
Observational Study of Dural Punctures
Pain as Disease and Illness: Part Two
Practice Patterns of Clinicians Treating Vulvar Pain
Share the Risk Model
Treating Sports-related Injury and Pain with Light Therapy
Using Topiramate in the Treatment of Migraine

Observational Study of Dural Punctures

A retrospective review of the rate of dural punctures as a complication of lumbar epidural steroid injection—with and without guiding fluoroscopy.

Epidural steroid injections are a routine treatment for low back pain despite the potential of dural puncture. The use of fluoroscopy remains controversial despite evidence that blind injections frequently are not placed correctly. Many studies report the technique and efficacy of steroid injections with no true consensus on either method.1,2 Despite the method used, complications can occur, including puncture of the dura, intravascular injection, nerve damage, hematoma, neurological compromise, and infection. Some of these complications are rare, but can be life threatening. Treatment studies published on epidural injections are not consistent in reporting complications during the epidural procedure.

Current opinion is mixed about the necessity of using fluoroscopy when performing epidural injections.3-5 Studies of fluoroscopic use during epidural injections have focused on correct needle placement in relationship to the site of pathology.5-9 These studies all support the use of fluoroscopy, citing needle placement accuracy rates of 98-99% when using fluoroscopy versus 48% to 62% without fluoroscopy. To our knowledge, no studies have addressed the number of dural punctures acquired when performing fluoroscopic-guided lumbar epidural injections. The intent of this study was to determine if the use of fluoroscopy reduces the number of dural punctures during epidural lumbar injections and thereby increases the safety of epidural injections.


This retrospective study compared the number of dural punctures observed in 500 fluoroscopically-guided lumbar epidural injections to the number reported in literature for non-fluoroscopically-guided lumbar procedures. Approval for this study was given by the Washington University Medical Center Humans Studies Committee. Medline and Embase was searched for the period of 1966 to 2006 using the following key words: lumbar, injections, steroids, epidural, sciatica, dural punctures, and complications. Citations within articles retrieved from these searches were also reviewed. Studies, used to compare dural puncture incidences were selected using the following criteria:

1) the study reported only lumbar epidural injections, 2) the study methodology clearly indicated that fluoroscopy had not been used to guide the needle placement, and 3) the study reported the number of dural punctures obtained in a specified number of injections.

Table 1: Dural Punctures Reported in Lumbar Injections*
Study or Manuscript Procedure to Locate Epidural Space Patient Diagnosis Number of Patients # of Dural Punctures Dural Punc.Rate (%)
Fredman (1999)7 Loss of Resistance using air Failed Back Surgery Syndrome 50 4 6.0
Carette (1997)10 Barry and Kendall procedure (hanging drop with loss of resistance) Sciatica 168 2 1.2
Rocco (1989)11 Two person technique using a saline extension tube as a manometer Post Laminectomy 24 1 4.2
Ridley (1988)12 Barry and Kendall procedure Sciatica 21 2 9.5
Klenerman (1984)13 Not specified Sciatica 52 1 2.0
Dilke (1973)14 Barry and Kendall procedure Sciatica–lumbar disc disease 51 6 11.8
Warr (1972)15 Not specified Sciatica or lumbago lumbar disc derangement spondylosis, spondylolisthesis and persistent/recurred symptoms after laminectomy 500 7 1.4
Burn (1970)16 Not specified Sciatica due to prolapse of intervertebral discs or spondylitis or spondylolisthesis 138 3 2.2
*All studies reviewed did not use fluoroscopic guidance of the needle. Fredman used fluoroscopy to check needle placement after insertion.
Table 2. Demographics for all 202 Patients in the Study
Sex: Male 93 46%
  Female 109 54%
Age:   57 (±15)
Radiographically Defined Diagnosis for Back Pain*:
  DDD 80 21%
  Herniated Disc 37 10%
  Protruding Disc 89 23%
  Spinal Stenosis 69 18%
  Spondylolisthesis 13 3%
  Spondylosis 27 7%
  Other** 70 18%
Patients with Previous Back Surgery:
    53 26%
*Only 179 of the 202 patients had a radiographically confirmed diagnosis for back pain. Several of the subjects had more than one diagnosis. **Other diagnoses include: Listhesis, fusion, degenerative polyneuropathy, scoliosis, hypertrophy, Schmorl’s Node, arthritis, osteoarthritis, arachnoiditis, facet arthropathy, radiculopathy, sclerosis, lumbago, disc desiccation, diskectomy, compression fractures, ankylosis, facet disease, myelopathy, levoscoliosis, Tarlov cyst.

Using these criteria, eight studies were reviewed and categorized according to: 1) the medical condition for which the injections were given, 2) the number of patients reported, and 3) the number of dural punctures reported.7,10-16

The charts of 202 patients receiving lumbar epidural steroid injections in the author’s community-based, academic-affiliated pain management center were examined. Procedure, nursing, and progress notes from five hundred epidural injection procedures were reviewed. The patient’s sex, age, number of injections, diagnosis, and complications noted — during the procedure and in the recovery area— were recorded.

The same basic epidural procedure was used for all 500 flouroscopically-guided injections. Initially, a fluoroscopic exam focused on the level being treated or, in the case of a general injection, identified the widest interlaminar space between L3 and S1. Epidural steroid injections were done on an outpatient basis with strict adherence to aseptic technique. An initial fluoroscopic picture was taken of the intended area of injection. Two cc of 1% lidocaine was placed subcutaneously along the intended path. After one minute, a 17-gauge needle was advanced through the skin. Fluoroscopic imaging was repeated every 1 to 2 cm to verify the desired alignment. The stylet was removed once the needle tip had encountered an increased resistance characteristic of the ligamentum flavum. Using 3 cc of air in a glass syringe, the needle was advanced while the plunger was balloted. Once a change of resistance occurred and the needle was considered to be in the epidural space, fluoroscopy (anterior, posterior, and lateral) was repeated. Omnipaque, ¼ to ½ cc, was injected if there was any uncertainty of the needle location. Correct placement was additionally checked by a negative aspiration of blood or CSF. Medication—typically 2 cc 1% lidocaine, 1 cc (80 mg) depomedrol, and 3 cc normal saline—was then injected.

Table 3. Demographics of Patients with Dural Punctures (7 dural punctures in 6 individuals*)
Sex: Male 1 17%
  Female 5 83%
Age: 64 (±11)    
Diagnosis: Protruding disc 3  
  Herniated Disc 1 17%
  Other 3 50%
  History of past back surgery: 1 17%
*A 67-yr-old female had two dural taps on two separate occasions. The woman had a radiographic diagnosis of protruding disc and levoscoliosis with no history of back surgery.


Historical rates: The dural puncture rate reported in the eight studies that were reviewed ranged from 1.2% to 11.8%. The number of injections reported in the articles varied from 21 to 500. One study indicated that fluoroscopy had been used to check for needle placement after a blind stick.7 However, none of the studies reported the use of intermittent fluoroscopy to guide the movement of the needle during the injection process. The results of the studies reviewed are summarized in Table 1.

Rate observed by investigator: Seven dural punctures, a rate of 1.4%, occurred in the 500 procedures reviewed. One patient received 2 dural punctures; therefore the 7 dural punctures corresponded to 6 individuals. Radiographic confirmation of a potential diagnosis responsible for the pain was available for 179 of the 202 subjects. Several subjects had more than one diagnosis. The demographic information for all 202 subjects is presented in Table 2. The demographics for the six patients receiving the dural puncture are presented in Tables 3 and 4.

A formal analysis was not conducted comparing the dural puncture rate in the author’s practice to the rates reported in the literature. The studies reviewed provided limited information about the procedure used to identify the epidural space, the variation in patient population, the variation in sample size, and the varying level of expertise among the physicians. Further analysis would not have provided a significant mean rate of dural punctures to allow a statistically sound comparison.


The use of fluoroscopy remains controversial despite evidence that blind injections frequently are not placed correctly.6 Most studies of epidural injections focus on the caudal approach and report dural puncture rates from 0% to 2.6%.17-20 The caudal approach usually has a decreased risk for dural puncture by the nature of the anatomy of the sacrum. In addition, many of the epidural injection studies report obstetric spinal block procedures that are usually performed in patients lacking significant degenerative processes. Lumbar epidural steroid injections frequently are given to patients with abnormal spinal structures that have resulted from surgery or degenerative diseases, thus making needle placement more difficult.

In this study, we found the dural puncture rate during fluoroscopically guided lumbar epidural injections to be 1.4% compared to 1.2% to 11.8% in the studies reviewed of non-fluoroscopically-guided lumbar injections. The Warr et al. and Carette et al. studies most closely resemble our study. Warr reviewed 500 epidural injections and observed a 1.4% dural puncture rate and Carette reviewed 168 epidural injections and observed a 1.2% dural puncture rate. Both studies had similar patient populations to the current study based on age and disease states. However, neither Warr nor Carette used guided fluoroscopy for needle placement.10,15

Three possible explanations for the wide variance in dural punctures reported in the literature are presented. First, the experience of the physician performing the injection affects the placement of the needle in the epidural space. In a study of caudal epidural steroid injections performed by radiologists with varying amounts of experience, Renfrew reported that more experienced physicians were able to place the needle in the epidural space 14% more often than inexperienced physicians (47.7% vs. 61.7%). In Renfrew’s study, only one dural puncture out of 328 procedures was noted; however, dural punctures are rarer with the caudal approach.9 Second, various spinal characteristics of the patient population being examined can complicate the injection procedure. Fredman et al. studied epidural steroid injections in patients with failed back surgery and suggested that local tissue trauma with fibrosis and adhesions from surgery might complicate the placement of the epidural needle.7 Third, the technique utilized by the physician to locate the epidural space also plays a role in complications encountered. Location of the epidural space is done by one of three methods: tactile, visual, and a modified tactile procedure using a mechanical device.21 The tactile method appears to be the most commonly used although variation in the use of saline or air added an additional variation to this method. It has not been established whether one method is superior to another and no clinical trials support the safety of one method over another.

The 500 procedures reviewed in this study were all completed by an experienced pain management anesthesiologist performing an estimated average of 100 lumbar epidural steroid injections per month over 7 years. The spinal problems of the patient population varied with the largest number being degenerative disc, protruding disc, or herniated disc. The physician performing the injections used the tactile method together with fluoroscopy for locating the epidural space.

In comparing the demographics of the six individuals experiencing the dural punctures to the individuals that did not receive dural punctures, those in the dural puncture group were older (64 vs. 57 years) and were predominately female (83% vs. 54% female). In this limited number of dural punctures, the diagnosis, the number of back surgeries, and the number of previous injections did not reveal any striking differences. Statistically, the numbers are not large enough to draw any conclusions.

There are limitations to this study. First, additional dural punctures may have occurred and were not recognized due to the lack of a positive aspiration or positive symptoms following the injection. The data collected from the investigator’s practice was limited to the time of injection and the time spent in the recovery area. Symptoms related to dural punctures, such as headaches, can occur after the patients have been discharged. For the review of dural punctures in the literature, the criteria for determining a dural puncture, other than aspiration, was not defined. Second, the studies reviewed for this paper were several years old. Techniques and procedures have advanced over time. Third, studies are not frequently designed to examine the dural puncture rate and many studies do not report dural punctures. After all, dural punctures represent a medical error—albeit often unavoidable—and so physicians may be reluctant to bring attention to the punctures.21

Table 4. Individual demographics of Dural Puncture Patients
Patient Number 1 2 3 4 5 6
Sex F F F F F M
Age 34 47 75 57 64 76
Diagnosis* DP PD PD,CF HD PD PD,LS
Number of Prior Back surgeries 1 0 0 0 0 0
Total number of injections 5 2 3 3 3 3
Injection number at which puncture occurred 4 1 2 1 1&3 3
* The diagnosis listed above include degenerative polyneuropathy (DP), protruding disc (PD), compression fracture (CF), herniated disc (HD) and levoscoliosis (LS).


This retrospective chart review determined the rate of dural punctures obtained by an experienced anesthesiologist performing fluoroscopic-guided lumbar steroid injections and compared it to the rate reported in literature of dural punctures obtained during non-fluoroscopic injections. Results indicate that, in the author’s practice, the rate of dural punctures during fluoroscopic-guided epidural injections was essentially the same as the lowest reported rate for non-fluoroscopic injections noted in other published studies.

The author has found that physician experience, technique, equipment, and patient population confound any retrospective study evaluating the need for fluoroscopy. Controlled studies comparing fluoroscopic versus non-fluoroscopic guidance of needle placement are needed to fully evaluate the need for fluoroscopy during epidural steroid injections.

Last updated on: December 20, 2011
close X