ORIGINAL COMMUNICATION

Anatomic Consideration of Caudal EpiduralSpace: A Cadaver Study

ANJALI AGGARWAL,1* HARJEET KAUR,1 YATINDRA K. BATRA,2 ADITYA K. AGGARWAL,3

SUBRAMANYAM RAJEEV,2 AND DAISY SAHNI1

1Department of Anatomy, Postgraduate Institute of Medical Education and Research,Chandigarh, India

2Department of Anaesthesia, Postgraduate Institute of Medical Education and Research,Chandigarh, India

3Department of Orthopaedic Surgery, Postgraduate Institute of Medical Education and Research,Chandigarh, India

The knowledge of sacral hiatus anatomy is imperative in clinical situationsrequiring caudal epidural block for various diagnostic and therapeutic proce-dures of the lumbosacral spine to avoid failure and dural injury. In this study,a detailed anatomic study of the sacral region was carried out on 49 male adultIndian cadavers. Dorsal surface of sacral region was dissected to study sacralcornua, sacral hiatus, and the dimensions of triangle formed by the right andleft posterosuperior iliac spines with apex of the hiatus. Midsagittal sectionswere subjected for various anatomical measurements. The angle of needleinsertion and the depth of caudal space were noted. Cornu was not palpablebilaterally in 7 (14.3%) and palpable unilaterally in 12 (24.5%) specimens.Mean (standard deviation) distance between apex of hiatus and coccyx tip was57.5 (8.7) mm and length of sacrococcygeal ligament was 34.2 (7.4) mm. Thedimensions of the triangle were found to be interchangeable in 25 cadavers.Once the needle is introduced into the canal after penetrating the sacrococcy-geal ligament, it should not be advanced >5 mm to prevent dural puncture.The level of maximum curvature of sacrum was S3 in 34 (69.4%) of cases.The dural sac was found to terminate at S2 in 41 (83.6%). The mean (SD)angle of depression of the needle was 65.7 (5.5) (range 58–78). The measure-ments described for the identification of the sacral hiatus, optimal angle ofdepression, and depth of the needle may improve the safety and reliability of acaudal epidural block. Clin. Anat. 22:730–737, 2009. VVC 2009 Wiley-Liss, Inc.

Key words: epidural space; sacrum; anesthesia; caudal; anatomy

INTRODUCTION

Caudal epidural block is widely used for intraoper-ative and postoperative analgesia as well as variousdiagnostic and therapeutic procedures on lower backby the anesthesiologists and orthopedic surgeons.The knowledge of caudal anatomy can have implica-tions not only for the anesthesiologist but also forthe radiologists, neurologists, and surgeons operat-ing on the spine (Srijit and Shipra, 2007). Prerequi-site for administering a successful caudal blockbegins with correct identification of sacral hiatus.Subsequently, needle is introduced into the sacral

canal through sacral hiatus after puncturing sacro-coccygeal ligament. The correct angle of depression

*Correspondence to: Anjali Aggarwal, Assistant Professor,Department of Anatomy, Postgraduate Institute of Medical Educa-tion and Research, Chandigarh, India or #123-C, Old Type V,Sector 24A, Chandigarh, 160012, India.E-mail: anjli_doc@yahoo.com or agg_adityadoc@yahoo.co.in

Received 27 February 2009; Revised 7 June 2009; Accepted 15June 2009

Published online 27 July 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ca.20832

VVC 2009 Wiley-Liss, Inc.

Clinical Anatomy 22:730–737 (2009)

of needle so as to lie along the axis of the sacralcanal is essential for successful caudal block. Theknowledge of the level of termination of dural sac isrequired to prevent injury to dura or surroundingstructures. The anatomic variations may contributeto failure of caudal epidural block. Crighton et al.(1997) in their study of anatomy of adult caudalcanal using magnetic resonance imaging suggestedthat successful entry into the caudal canal was bestachieved through upper third of the sacrococcygealligament. The purpose of this study was to definethe anatomy of sacral hiatus and to determine angleof needle insertion that would increase the likelihoodof entry of the needle into sacral hiatus thus enhanc-ing the success rate of caudal epidural block. Weassessed interchangeability of measurements ofboundaries of the triangle which formed the anatom-ical basis for performance of caudal block.

MATERIALS AND METHODS

Institutional review board approval was obtained.The study was carried out on 49 male adult Indiancadavers in the department of Anatomy Postgradu-ate Institute of Medical Education and Research,Chandigarh, India. The cadavers were placed in theprone position. Dorsal surface of sacral region wasdissected, adjacent soft tissue ligaments and muscu-lar attachments were removed. The cadaver withundamaged sacral bone was considered for furtheranalysis. Measurements were performed by first twoauthors separately using divider, thread, protractor,and sliding vernier caliper to the accuracy of 0.02mm. Each item was measured twice and averagedby two authors independently. Mean of the total databy each investigator was considered.

Measurements on the DorsalSurface (Fig. 1)

Sacral cornu and posterosuperior iliac spines wereused as bony landmarks for measurements. Caseswith palpable cornu were noted. Most prominentpoints on right and left posterosuperior iliac spineswere marked with a marker and apex of sacral hiatuswas felt manually. A triangle was constituted by join-ing these three points. To locate the sacral hiatus,measurements were performed as following:

1. Interspinous line formed by joining right andleft posterior superior iliac spines representedbase of the triangle (a).

2. Right side of the triangle was made by joiningright posterior superior iliac spine with apex ofhiatus (b).

3. Left side of the triangle was made by joiningleft posterior superior iliac spine with apex ofhiatus (c).

4. Intercornual distance between inner surfacesof tips of sacral cornua (d).

Measurements on Midsagittal Section(Figs. 2 and 3)

Sacrum was subsequently sectioned in midsagittalplane and following parameters were measured:

Fig. 1. Parameters measured on dorsal surface ofsacral region: (a) Interspinous line distance betweenright and left posterior superior iliac spines, (b) rightposterior superior iliac spines and apex of sacral hiatus,(c) left posterior superior iliac spine and apex of sacralhiatus, and (d) intercornual distance.

Fig. 2. Parameters measured in sagittal section:(e) Termination of dura and apex of hiatus, (f) length ofsacrococcygeal membrane, (g) distance between apexof hiatus and tip of coccyx, (h) anteroposterior depth ofepidural space at the apex, and (i) angle of depressionof needle.

731Anatomic Study of Caudal Epidural Space

1. Straight distance between end of dural sac andapex of hiatus (e)

2. Length of sacrococcygeal ligament coveringcaudal space (f)

3. Curved distance between tip of coccyx andapex of hiatus (g) (with a thread)

4. Anteroposterior depth of caudal space at theapex of sacral hiatus (h).

5. Angle of depression of the needle into the sacralcanal (i) was measured to determine the suc-cess of a caudal epidural block. For measuringthis angle, needle was first inserted through theskin at right angle to sacrococcygeal ligament.Then, it was depressed by such an angle so thatit was in the plane of sacral canal. This angle ofneedle depression lies between the two lines:(1) perpendicular to the sacrococcygeal liga-ment and (2) along the axis of sacral canal andis measured by a protractor placed against thecut midsagittal surface. This method has beendepicted in Figure 2 (angle of needle depressionis indicated by ‘‘i’’) of the manuscript.

Level of termination of dural sac (Fig. 4) and max-imum curvature of sacrum in relation to adjacentsacral vertebral body were also noted.

Statistical Analysis

Post-hoc power analysis was performed with cor-relation coefficient of the measurements of threesides of the triangle. Highest sample size obtainedwas considered based on correlation of two amongthe three techniques. A sample size of 25 was foundto be adequate to demonstrate interchangeability ofthree measurements with an alpha of 0.05 and 90%power. Bland–Altman and mountain plots (foldedempirical cumulative distribution plot) were used toassess interchangeability of the measurementsobtained (Bland and Altman, 1986; Krouwer andMonti, 1995). Plots were assessed with Medcalc(Medcalc software, Belgium). Bias and precision

were calculated from Bland–Altman plots. Data wereexpressed as mean (SD) and range. The interob-server reliability and intraobserver reproducibilitywere analyzed by Kappa coefficient.

Fig. 3. Shows method of measuring curved dis-tance between apex of sacral hiatus and tip of coccyxusing thread.

Fig. 4. Midsagittal sections of sacral region showingtermination of dural sac at: (a) level of S1, (b) level ofS2, and (c) lower border of S3.

TABLE 1. Measurements on Dorsal Surface ofSacral Region

Parameter Minimum Maximum Mean SD

Base ofthe triangle

51.95 86.86 70.1 7.8

Right borderof the triangle

45.07 86.86 66.0 9.6

Left borderof the triangle

47.24 88.09 65.3 9.8

Intercornualdistance

3.40 14.00 7.9 2.6

732 Aggarwal et al.

RESULTS

Parameters Studied on the Dorsal Surface(Table 1)

Mean (SD) interspinous measurement (base ofthe triangle) was 70.1 (7.8) mm. Mean (SD) dis-tance measured from right posterosuperior iliacspine (right border of triangle) and left posterosupe-rior iliac spine (left border of triangle) was found tobe 66 (9.6) mm and 65.3 (9.8) mm, respectively.Cornu was not palpable bilaterally in 7 (14.3%)specimens, palpable bilaterally in 30 (61.2%), andunilaterally in 12 (24.5%) specimens. Mean (SD)intercornual distance was 7.9 (2.6) mm.

Measurements in Sagittal Section(Table 2)

Mean (SD) distance between apex of hiatusand coccyx tip was 57.5 (8.7) mm and length of

sacrococcygeal ligament was 34.2 (7.4) mm. Mean(SD) distance between apex of hiatus and dural sacwas 31.6 (11.8) mm. Level of maximum curvature ofsacrum was found to be S4 in 11 cadavers (22.5%),S3 in 34 (69.4%), between S3/S4 in one (2%), S2 in2 (4.1%), and between S2/S3 in one cadaver (2%).Level of termination of dura mater was found at S2in 41 (83.6%) (Fig. 4b), S3 in 4 (8.2%) (Fig. 4c),and S1 in 4 (8.2%) specimens.

Measurements With Implications forNeedle Insertion (Fig. 2)

Mean (SD) angle of depression of the needle was65.78 (5.58) (range: 58–788). Mean (SD) anteroposte-rior depth of caudal space at apex of the hiatus was 4(1.6) mm (range: 1–9.3 mm). Bland–Altman plotsshowed that measurements obtained on three sides ofthe triangle were interchangeable (Figs. 5–7). Biasand precision of the measurements was also accepta-ble (Table 1). A few outliers were observed in the

TABLE 2. Measurements in Sagittal Section

Parameter Minimum Maximum Mean SD

Curved distance between coccyx tip and apex of hiatus (mm) 38.27 81.76 57.5 8.7Straight length of sacrococcygeal ligament (mm) 18.25 48.96 34.2 7.4Straight distance between apex of hiatus and dural end (mm) 5.76 59.78 31.6 11.8A-P depth of caudal space at apex of hiatus (mm) 1.08 9.30 4.0 1.6Angle of depression of needle (8) 58 78 65.7 5.5

Fig. 5. Bland and Altman plot of data for base and right border of the trian-gle. The x-axis is the average of the two variables. The limits of agreement (2SD) were +19.4 and 11.4 as shown in dashed lines above and below zero.

733Anatomic Study of Caudal Epidural Space

scatter of plots, which suggested presence of extremevalues. Folded empirical cumulative plots showed thatcentral 95% of the data (peak of the plots) werecentered almost close to ‘‘0’’ suggesting low bias and‘‘tails’’ of the mountains were not widespreadsuggesting that intermeasurement variation wasminimal (Fig. 8). The interobserver reliability andintraobserver reproducibility were 0.82 and 0.87,respectively.

DISCUSSION

Sacral hiatus forms one of the most importantlandmarks for performing caudal epidural block andfailure to appreciate the hiatus and other bony land-marks reduces the rate of a successful block in thepatients. This underscores importance of performinga detailed anatomic study on the sacra without sup-porting connective tissue structures (Sekiguchiet al., 2004), when compared with that performedon adult cadavers in this study. Most of the anatomi-cal studies reported in literature on adult sacral canalare >60 years old (Edwards and Hingson, 1942;Lanier et al., 1944; Trotter, 1947). This is anotherreason for conducting this study as there have beenanatomical changes in the population over the pastmore than five decades (Crighton et al., 1997).Although many studies have confirmed simplicity,reliability, and safety of caudal epidural block, signif-icant complications can ensue. In caudal epiduralblock, identification of sacral hiatus is very essential.

Sekiguchi et al. (2004) studied 92 isolated dry sacralbones to clarify anatomical variations of hiatus in theJapanese and defined existence of sacral cornua byits height being >3 mm. The cornu was absent in54% cases. This study was conducted on 49 cadav-ers and the authors tried to palpate tip of sacral cor-nua. In those specimens in whom tips of the cornuawere palpable, measurements were performedbetween their inner surfaces after removal of subcu-taneous tissues. The authors found some interestingdifferences in observations on Indian cadavers. Thecornu was not palpable bilaterally in 14.3% of casesand palpable only unilaterally in 24.5% of cases.This suggests that cornu cannot be relied upon forlocating sacral hiatus in 14% cases in Indian popula-tion when compared with 54% cases in Japanesepopulation. Intercornual distance needs to be suffi-cient enough for easy passage of needle into caudalspace. Sekiguchi et al. (2004) and Senoglu et al.(2005) found this distance to be 17.5 (3.2) mm and10.2 (0.35) mm, respectively, which was greaterthan 7.9 (2.6) mm observed in this study. Drawinginverted triangle by joining two posterosuperior iliacspines and apex of sacral hiatus is considered to bean important landmark for locating hiatus. This isclassically considered to be an equilateral triangle(Martin, 1987; Senoglu et al., 2005). We observedthat dimensions of all borders of triangle might notbe necessarily similar in an individual specimen. Theauthors have observed equilateral triangle in 51%cases only and large variation in the distances in49% cases. Hence, the authors believe that the

Fig. 6. Bland and Altman plot of data for base and left border of the triangle.The x-axis is the average of the two variables. The limits of agreement (2 SD)were +20.5 and �11 as shown in dashed lines above and below zero.

734 Aggarwal et al.

Fig. 7. Bland and Altman plot of data for right and left border of the triangle.The x-axis is the average of the two variables. The limits of agreement (2 SD)were +4 and �5.8 as shown in dashed lines above and below zero.

Fig. 8. Mountain plot comparing the interchangeability of the three measure-ments of the triangle shows that the central 95% of the data is centered close tozero. [Color figure can be viewed in the online issue, which is available atwww.interscience.wiley.com.]

735Anatomic Study of Caudal Epidural Space

equilateral triangle can not be fully relied upon forlocating sacral hiatus in all the cases. Thus, weattempted to confirm this variable statistically byassessing whether dimensions of the triangle wereinterchangeable or not. This was performed withBland–Altman plot, commonly used to test agree-ment of two variables and they correlated with eachother in 25 (51%) cadavers only. Use of correlationdoes not imply that there is good agreementbetween the variables. This plot is complementedwith the mountain plot. The scatter of the differencesobserved in between two dimensions were distrib-uted well within the limits of agreement (1.96 3 SD)between all the dimensions, suggesting that theywere interchangeable. Measurement of distancebetween apex of hiatus and tip of coccyx formsanother important anatomical parameter for locatingsacral hiatus. With the finger placed in natal cleftwith distal end at tip of coccyx, the sacral hiatus willbe at the level of proximal interphalangeal joint ofthe clinician (Martin, 1987). Mean (SD) value of thisdistance was 57.5 (8.7) mm. Mean anteroposteriordepth of the caudal space at apex of hiatus was 4mm. This has clinical implications for the provider ofthe caudal block for safe and effective caudal block.Sekiguchi et al. (2004) reported closed sacral canalin 3% of the studied population, thus rendering cau-dal epidural block impossible. Anteroposterior diame-ter of sacral canal was <2 mm in 1% of cases whichmight increase the chances of failed caudal blockdue to difficulty in needle insertion. In this study,diameter of sacral canal was <2 mm in three cases(n ¼ 49), that is, 6.1%. There was no incidence ofclosed sacral canal that was comparable with thestudy by Senoglu et al. (2005). For performing cau-dal block, needle is first inserted at right angle toskin surface, then depressed by such an angle that itlies along the axis of sacral canal. This angle is calledangle of depression of needle. This method is similarto that described by Martin (1987), which is quitedifferent from that used by Park et al. (2006). In thisstudy, mean value of this angle was found to be65.78 which was higher than 57.98 (mean) reportedby Crighton et al. (1997). The variations in the anglemay be attributed to racial diversity of the studiedpopulation. Park et al. (2006) found no significantcorrelation of optimal angle of insertion of needlewith age, height, weight, or body surface area. It isimportant to measure the straight distance betweensacral hiatus and end of dural sac which may help indeciding length of needle to be introduced into theepidural space so as to prevent dural injury. Senogluet al. (2005) have measured the distance from S2vertebral foramen to apex of sacral hiatus in dry sac-ral bones based on the presumption that the duralsac terminates at the level of S2 in all the cases andthey found mean (SD) to be 35.4 (10.4) mm. How-ever, the authors in this study observed dural sacterminating at different levels in different cadavers.The termination of dura occurred at S2 in >80% ofthe cases. What may be more important is the termi-nation of dura at S3 in about 8% when the clinicianattempting caudal epidural block at this level maycause dural puncture. Hence, in this study, the

authors measured the straight distance betweenapex of hiatus and end of dural sac, indicating theshortest distance between the two which was foundto be 31.6 (11.8) mm [mean (SD)]. Length of needleused for caudal block should be chosen carefully,longer the needle more are the chances of duralpuncture. According to the insertion length of nee-dle, the range of angle might be different. This mayprovide a useful data. Surprisingly in one of thecases, the distance was found to be as low as 5.76mm. Thus, the needle should be advanced carefullyfor only a few millimeters after penetrating the sac-rococcygeal ligament, to reduce chances of duralpuncture and other possible complications. Straightneedles can penetrate into bone. If the sacrum ismissed penetration of bowel can lead to serious com-plications. In the clinical arena, the curvature of thesacrum is solved by the R-X Coude needle that has acurved body. This is for better directional placementof site specific catheters in the lumbar epidural spacevia the sacral access. This may be useful in treat-ment such as the lysis of adhesions procedure that isdone world wide. Mean length of sacrococcygeal liga-ment in our study was 34.2 mm. This was higherthan mean length of 22.6 mm measured in MRIstudy conducted by Crighton et al. (1997). This maybe due to population or methodology differences.The authors observed that maximum curvature ofsacrum occurred at S3 in 64.7% of cases. This maybe important for surgeons operating on the spineand may guide in the management. In fact, a recentstudy showed that with the presence of a steep sac-ral slope C-arm guided reduction and cage insertionmethod is a reliable way of treating spondylolisthesis(Kim et al., 2008). Chen et al. (2004) claimed thatthe use of ultrasound to guide needle placement inthe caudal epidural space would increase successrate by 100%. Park et al. (2006) used ultrasoundimaging for estimating the optimal angle for needleinsertion and recommended its use for successfulcaudal epidural block in children. Fluoroscopy is con-sidered as the gold standard for determining accu-rate needle placement and for reducing the risk ofcomplications (Stitz and Sommer, 1999). The cost-effectiveness, time constraints, and availability oftrained personnel may limit the feasibility of ultra-sound or fluoroscopy particularly in the developingcountries. Thus, the importance of an accurateknowledge of such measurement is emphasized andsuch anatomical study carried out in cadavers maybe very useful in clinical situations.

We conclude that anatomy of the sacral canal isdiverse with racial differences. The previous studieshave been conducted on the dry sacral bones. Wehave carried out this study on the cadavers that pro-vides us the real-life situation, more close to clinicalsettings. We measured the curved distance from tipof coccyx to apex of sacral hiatus, which was an im-portant parameter for locating the sacral hiatus. Thishas not been previously reported in the literature tillnow. In our study, triangle was constructed by join-ing two posterior superior iliac spines with apex ofsacral hiatus when compared with Senoglu et al.(2005), who described the triangle formed between

736 Aggarwal et al.

superolateral sacral crest and apex of sacral hiatusin dry sacral bones. As equilateral triangle wasobserved in only 51% cases, hence, constructingsuch a triangle cannot be relied upon. Thus for locat-ing sacral hiatus, combination of landmarks ratherthan single should be used. We recommend the pal-pation of sacral cornua as the best landmark forlocating sacral hiatus. Estimation of curved distancefrom tip of coccyx to apex of sacral hiatus is anotherimportant measurement not reported in the litera-ture so far. Anteroposterior depth of sacral canal atthe apex of sacral hiatus is also an important param-eter because <2 mm depth may not allow the needleto enter the sacral canal. Once the needle isintroduced into the canal, it should not be advanced>5 mm to prevent dural puncture. The needle tipshould not be above the S-3 neural foramen as thedura may extend to that level. In failed back surgerypatients where there may be dural tears thefluid injected may dissect into the subdural andsubarachnoid spaces.

Angle of needle depression should be around 658to maximize the chance of success of caudal epiduralblock. The knowledge of angle of needle depressionand anteroposterior depth of caudal space can helpthe anesthesiologists in performing a safe and suc-cessful block. The measurements may be useful forradiologists and surgeons operating in this region. Inaddition, the measurements we obtained from thesacra can have clinical importance in other areasalso.

ACKNOWLEDGMENTS

The authors thank Mr. Vijay Kant Bakshi forartwork and photography work.

REFERENCES

Bland JM, Altman DG. 1986. Statistical methods for assessingagreement between two methods of clinical measurement. Lan-cet 1:307–310.

Chen CP, Tang SF, Hsu TC, Tsai WC, Liu HP, Chen MJ, Date E, LewHL. 2004. Ultrasound guidance in caudal epidural needle place-ment. Anesthesiology 101:181–184.

Crighton IM, Barry BP, Hobbs GJ. 1997. A study of the anatomy ofthe caudal space using magnetic resonance imaging. Br JAnaesth 78:391–395.

Edwards WB, Hingson RA. 1942. Continuous caudal anesthesia inobstetrics. Am J Surg 57:459–464.

Kim SY, Maeng DH, Lee SH, Jang JS. 2008. Anterior lumbarinterbody fusion for lumbosacral junction in steep sacral slope.J Spinal Disord Tech 21:33–38.

Krouwer JS, Monti KL. 1995. A simple, graphical method to evaluatelaboratory assays. Eur J Clin Chem Clin Biochem 33:525–527.

Lanier VS, McKnight HE, Trotter M. 1944. Caudal analgesia: Anexperimental and anatomical study. Am J Obst Gynec 47:633–641.

Martin LVH. 1987. Sacral epidural (caudal) block. In: Wildsmith JAW,Armitage EN, editors. Principles and Practice of Regional Aanaes-thesia. Edinburgh: Churchill Livingstone. p 103–110.

Park JH, Koo BN, Kim JY, Cho JE, Kim WO, Kil HK. 2006. Determina-tion of the optimal angle for needle insertion during caudal blockin children using ultrasound imaging. Anaesthesia 61:946–949.

Sekiguchi M, Yabuki S, Satoh K, Kikuchi S. 2004. An anatomic studyof the sacral hiatus: A basis for successful caudal epidural block.Clin J Pain 20:51–54.

Senoglu N, Senoglu M, Oksuz H, Gumusalan Y, Yuksel KZ, ZencirciB, Ezberci M, Kizilkanat E. 2005. Landmarks of the sacral hiatusfor caudal epidural block: An anatomical study. Br J Anaesth95:692–695.

Srijit D, Shipra P. 2007. Spina bifida with higher position of sacralhiatus: A case report with clinical implication. Bratisl Lek Listy108:467–469.

Stitz MY, Sommer HM. 1999. Accuracy of blind versus fluoroscopi-cally guided caudal epidural injection. Spine 24:1371–1376.

Trotter M. 1947. Variations of the sacral canal: Their significance inthe administration of caudal analgesia. Anesth Analg 26:192–202.

737Anatomic Study of Caudal Epidural Space