HlbamppcHuwHa

HeatfsHtdadno

FtAeLRAC©0d

C

Hysterosalpingography: Technique andApplications

Athanasios Chalazonitis, MD, PhD, MPH,a Ioanna Tzovara, MD,b

Fotios Laspas, MD, MSc,c Petros Porfyridis, MD,a Nikos Ptohis, MD, PhD,a

and Georgios Tsimitselis, MDd

dp

IHiHfirbcmta

s2etti

paomeivuoto

ysterosalpingography (HSG) remains an important radio-ogic procedure in the investigation of infertility and hasecome a commonly performed examination due to recentdvances of reproductive medicine. HSG demonstrates theorphology of the uterine cavity, the lumina, and theatency of the fallopian tubes. In this review article, weresent the technical parameters of the examination, indi-ations, contraindications, and possible complications ofSG. We also illustrate a variety of abnormalities of theterus and fallopian tubes that can be detected accuratelyith HSG. We believe that, with the increased demand forSG, radiologists should be familiar with HSG techniquend the interpretation of HSG images.

ysterosalpingography (HSG) is the radiographicvaluation of the uterine cavity and fallopian tubesfter the administration of a radio-opaque mediumhrough the cervical canal. The first HSG was per-ormed in 1910 and was considered to be the firstpecial radiologic procedure. A properly performedSG can detect the contour of the uterine cavity and

he width of the cervical canal. Further contrast me-ium injection will outline the cornua isthmic andmpullary portions of the tubes, and will show theegree of spillage. If a properly performed HSG showso uterine cavity abnormality, it is very unlikely thatther modalities would do so.1 Although this proce-

rom the aDepartment of Radiology, General Hospital of Athens Hippocra-io, Athens, Greece; bDepartment of Radiology, IASO General Hospital,thens, Greece; cDepartment of Radiology, Elpis General Hospital, Ath-

ns, Greece; and dDepartment of Radiology, Larissa University Hospital,arissa, Greece.eprint requests: Fotios Laspas, MD, MSc, Thisseos 29, 15234 Halandri,thens, Greece. E-mail: fotisdimi@yahoo.gr.urr Probl Diagn Radiol 2009;38:199-205.2009 Mosby, Inc. All rights reserved.

363-0188/2009/$36.00 � 0

toi:10.1067/j.cpradiol.2008.02.003

urr Probl Diagn Radiol, September/October 2009

ure is considered diagnostic, there may also be aossible therapeutic benefit from the flushing effect.2,3

ndications and ContraindicationsSG is used predominantly in the evaluation of infertil-

ty.4 Despite the arrival of newer imaging modalities,SG still remains the best procedure to image the

allopian tubes.5 Although evaluating feminine infertil-ty, with or without the presence of repeated miscar-iages, is the main indication for this method, it can alsoe used in other cases, such as pain in the pelvis tract,ongenital or anatomic abnormalities, anomalies of theenstrual cycle, and abnormal menses. Also, it is some-

imes used as a preoperative control for women who arebout to have uterine or tubal surgery.6

Soares and coworkers showed that HSG had aensitivity of 58% and a positive-predictive value of8.6% for polypoid lesions, and a sensitivity of 0% forndometrial hyperplasia. The same study showed HSGo have a sensitivity of 44.4% for uterine malforma-ions, and a sensitivity of 75% for the detection ofntrauterine adhesions.7

The main contraindication of the examination isossible pregnancy. This contraindication can bevoided by performing the examination before thevulation phase, between the 7th to 10th day of theenstrual cycle.6 Because of the scattering risk, the

xamination should be avoided when there is activentrapelvic inflammation. Another contraindication isaginal or uterine bleeding because of the risk ofnrestrained bleeding, which could lead to transfusionr surgical recovery procedures. Finally, the examina-ion should not be performed in cases of severe cardiacr renal deficiency, or in cases of recent uterine or

ubal surgery.8

199

T

P

mewaatcssnn

amAtpsr1

ifi

C

otadsoopewbtoaubds

cb

C

udlccptmtpo

R

�cctAtestifsetetBc

CTae

2

echnique

atient PreparationThe procedure is performed in the first half of the

enstrual cycle following cessation of bleeding. Thendometrium is thin during this proliferative phase,hich facilitates better image interpretation and should

lso ensure that there is no pregnancy. The patient issked to refrain from unprotected sexual intercourse fromhe date of her period until after the investigation to beertain there is no risk of pregnancy. Examination in theecond half of the cycle is avoided because the thickenedecretory-phase endometrium increases the risk of ve-ous intravasation and may cause a false-positive diag-osis of cornual occlusion.9

Antibiotics might be required 1 day before and forfew days after the examination if previous inflam-ations are present in the patient’s clinical history.ntibiotics are required after the examination when

he maneuvers are fairly sanguineous or if the fallo-ian tubes present a certain degree of dilation. Theuggested antibiotic regimen is metronidazole 1 gectally at the time of the procedure, plus doxycycline00 mg twice daily for 7 days.10

Steroid (prednisolone) premedication is prescribedn asthmatics when intravenous contrast is used; there-ore, it is reasonable to do the same for HSG becausentravasation is also possible from this procedure.3

atheterization Technique4,11,12

For the catheterization technique, the patient is placedn the fluoroscopic machine in a gynecologic examina-ion position. After cleaning the external genital area withntiseptic solution, the vagina is dilated by a gynecologicilator. The cervix is localized and cleansed with iodineolution. Afterward, the uterine cervix is straightened byne (at the 12 o’clock position) or two (at the 9 and 3’clock positions) surgical forceps exercising a degree ofulling. Next, the outside uterine cervix ostium is cath-terized. The catheterization can be performed in twoays. In the authors’ country, a salpinographer with aell-shaped end (diameter depends on the case) is pushedhrough the vagina and fits in the external uterine cervixstium. In the second technique, the salpingographer hasplastic cup-shaped end that is fitted to the external

terine cervix ostium, creating a void phenomenon. Inoth techniques, there is a syringe with iodinated hy-rosoluble contrast medium at the other end of the

alpingographer. The vagina dilator is taken off after

00

atheterization of the external uteri cervix ostium andefore administration of the contrast medium.

ontrast MediaIn the past, oil-soluble contrast media were mainly

sed. Today, we use all available iodinated hy-rosoluble contrast media. According to internationaliterature, the use of oil-soluble contrast media in-reases the pregnancy rate and contributes to a de-rease in conception time after the salpingography iserformed.13,14 However, Spring and coworkers foundhat there is no evidence that the choice of the contrastaterial affects the rate of term pregnancy. Moreover,

hey reported that oil-soluble contrast media mayromote granulomatous inflammation in the presencef obstructed or inflamed fallopian tubes.15

adiological ViewsOne conventional radiograph of the pelvis (on a 2430 cm radiologic film) is necessary before the

ontrast medium is administrated into the uterineavity so that possible intrapyelic masses or calcifica-ions will not complicate interpretation of the images.

metallic marker is placed over one side of the pelviso indicate the right or left side of the patient. Next, thexamination is performed under fluoroscopic controlo that radiographs can be taken during the filling ofhe uterine cavity (usually 2-3 cm3 of contrast mediums sufficient) and again during the filling of theallopian tubes. Finally, after the removal of thealpingographer, we radiographically check the pres-nce of contrast medium in the peritoneal cavity. Theotal amount of injected contrast medium should notxceed 10 mL. Additional spot radiographs are ob-ained to document any abnormality that is seen.efore the first radiograph, we also fluoroscopicallyheck the reflux of the contrast medium.16

omplications4,17-19

he two most common complications of HSG are painnd infection. These and other complications and sideffects are summarized below.

● Uterine contractions and discomfort due to theintroduction of contrast medium into the uterinecavity: The most common type of pain referencedis subabdominal colic caused by dilation of theuterine cavity. A more diffuse pain, caused by

irritation of the peritoneum due to the contrast

Curr Probl Diagn Radiol, September/October 2009

NOttpobbofsum(uadmscnc

CC

Fm

FFt

C

medium, has also been reported. Pain can beminimized by slowly injecting the contrast me-dium and using isosmolar contrast agents.

● Postprocedural infection: Spreading and general-ization of intrapyelic inflammation may happenin cases of chronic inflammation and hydrosal-pinges, or after severe uterine injury caused bythe examination maneuver.

● Vasovagal reaction: A possible reaction to ma-nipulation of the cervix or inflation of a conclu-sion balloon in the cervical canal.

● Traumatic elevation of endometrium by the in-serted cannula: A complication which does notcause significant consequences.

● Uterine perforation and tubal rupture: Thesecomplications are very rare.

● Venous or lymphatic intravasation of contrastmedia: With a water-based contrast medium thereis no adverse effect on the patient, but it can makeinterpretation of the image difficult. It occursmore commonly in the presence of fibroids ortubal obstruction. Extravasation of the contrastmedium (Fig 1) could occur if the contrastmedium is administered too quickly, if the endo-metrium is injured during the catheterization, orif the examination is performed during menstru-ation. Extravasation is also possible when com-mon or special inflammations of the endome-trium are present due to the intercourse ratebetween the uterine vein and the ovarian veins.

● Allergic reaction to contrast media: Such a reac-

IG 1. Extravasations of the contrast medium. Presence of contrastedium in the peritoneum.

tion is very uncommon with the use of the j

urr Probl Diagn Radiol, September/October 2009

low-osmolar nonionic contrast agents currentlyavailable.

● Radiation exposure to the ovaries: Exposure isminimal and can be reduced if the proper tech-nique is utilized.

ormal Findingsn face radiographs, the uterine cavity has a normal

rigonal shape and the apex of the triangle corresponds tohe isthmus, which is nearly 3.7 cm wide. The apex isointed downwards and connected to the internal ostiumf the cervix uteri, which is 2.5 cm in total length. Thease of triangular uterine cavity is the fundus, which cane concave, flattened, or slightly convex. On both sidesf its base, in the area of the lateral horns, the twoallopian tubes are drowned. The fallopian tubes areeparated into three segments: isthmus (attached to theterus, not imaged in several cases), ampullary (in theiddle, the longest and widest segment), and bell-shaped

to the distal end). There are two ostiums: the internal orterine, and the external or abdominal (Fig 2). From thebdominal ostium, the contrast medium disperses andiffuses into the peritoneal cavity. Remaining contrastedium in the furrows of the peritoneum can be ob-

erved up to 3 hours after administration. Very often, theontrast medium in the rectouterine pouch of the perito-eum (Douglas’ space) can demonstrate the profile of theoordinate ovary.8

ongenital Uterus Anomaliesongenital uterus anomalies are caused by incomplete

IG 2. Normal hysterosalpingography. Uterus in right inclination.ull-length drawing of the vagina, the uterus cavity, and the fallopianubes.

unction of the paramesonephric ducts (Muller ducts), or

201

bbtpanidrabnbiw3

ahpdiplhaAp

Ml

C

C

C

C

C

C

C

hs

Fa

Fc

2

y the nonabsorption of the diaphragm, which is locatedetween ducts during the development of the uterus inhe 18th week of the pregnancy.19 The true incidence andrevalence of mullerian duct anomalies are difficult tossess.20 Examination of different patient populations,onstandardized classification systems, and differencesn diagnostic data acquisition has resulted in widelyisparate estimates, with a reported prevalence thatanges from 0.16 to 10%.21 As a result of selection bias,prevalence of 8 to 10% has been reported in women

eing evaluated with HSG because of recurrent preg-ancy loss.22 The overall data suggest that the prevalenceoth in women with normal fertility and in women withnfertility is approximately 1%, and the prevalence inomen with repeated pregnancy loss is approximately%.23

While the majority of women with mullerian ductnomalies have little problem conceiving, they haveigher associated rates of spontaneous abortion,remature delivery, and abnormal fetal position andystocia at delivery. Most studies report an approx-mate frequency of 25% for associated reproductiveroblems, compared with 10% in the general popu-ation. Primary infertility in these women usuallyas an extra uterine cause and is not generallyttributable to mullerian duct anomalies alone.24

dditionally, cervical incompetence has been re-orted to be associated with these anomalies.25

According to the American Society of Reproductiveedicine,26 there are seven different classes of mul-

erian duct anomalies:

lass I: Segmental agenesis or variable degrees ofuterovaginal hypoplasia. The anomaly can be de-tected, because of the amenorrhea, before HSG isperformed.

lass II: Unicornuate uteri (Fig 3) that representpartial or complete unilateral hypoplasia. In rarecases of degeneration of the mesonephric duct, theuterine cavity appears monocular when imaged,placed right or left of the middle line. The unicor-nuate uterus contacts only the coordinate fallopiantube.

lass III: Didelphys uterus. This is a rare abnormalitythat results from complete nonfusion of the mulle-rian ducts, and includes the duplication of theuterine cavity, cervix neck, and vagina. Rarely, this

uterus has a single vagina (Fig 4). c

02

lass IV: Bicornuate uterus (Fig 5) that demonstratesincomplete fusion of the superior segments of theuterovaginal canal. The uterine cavity is divided intwo; each half has a narrow-length shape andstands apart from the other.

lass V: Septate uteri that represent partial or com-plete nonresorption of the uterovaginal septum.

lass VI: Arcuate uterus (Fig 6) resulting from nearlycomplete resorption of the septum. Along with theprevious anomaly, these are the most commoncongenital anomalies (50%) in cases detectingfemale infertility.

lass VII: Anomalies that comprise sequelae of inutero diethyloestradiol exposure.

Another congenital anomaly, caused by inadequateormonic stimulation as a fetus, is small uterine cavityize with normal vaginal length (Fig 7). This is a

IG 3. Unicornous uterus. Hysterosalpingography shows opacification ofsingle right uterine horn. A single fallopian tube is also visualized.

IG 4. Didelphys uterus. Hysterosalpingography shows two uterineavities, two cervices, and one single vagina.

ommon finding in cases of female infertility.

Curr Probl Diagn Radiol, September/October 2009

AF

sstodvvfw

E

efis

miSo

I

itoaimdmtc

U

d

Ff

Fs

F

Fb

C

bnormal Findingsibromyomas

While fibromas are diagnosed by suprapubic ultra-ound, submucosa fibromyomas (Fig 8) are imaged asmooth filling defects in the uterine cavity. Differen-ial diagnosis must be made from endometrial polypsr possible pregnancy. Small intramural fibromyomaso not distort the endometrial cavity and are notisualized on HSG. Subserous fibromyomas can pro-oke smooth filling defects or smooth repression of theallopian tubes only if they are located in the lateralalls of the uterus.

ndometrial PolypsEndometrial polyps are focal overgrowths of the

ndometrium. They usually manifest as well-definedlling defects and are best seen during the early filling

IG 5. Bicornate uterus. Spot radiograph shows two uterine horns. Theallopian tubes are also visualized at this imaging stage.

IG 6. Arcuate uterus. Hysterosalpingography demonstrates a depres-ion of the uterine fundus, compatible with an arcuate uterus.

tage. Small polyps may be obscured when contrast m

urr Probl Diagn Radiol, September/October 2009

aterial completely fills the uterine cavity and may bendistinguishable from a small submucosal myoma.onohysterography has become the preferred methodf imaging endometrial polyps.4

nternal Endometriosis (Adenomyosis)Adenomyosis is caused by the presence of ectopic

slets of active endometrium in the muscularis wall ofhe uterus. It is usually imaged as a pointed projectionf 2 to 3 mm length, perpendicular to the uterine wallfter contrast medium administration. Rarely, this ismaged as a sack-shaped projection filled by contrastedium, 4 mm to 1 cm in length (Fig 9). Differential

iagnosis should include the hyperplasia of the endo-etrium and the entrance of the contrast medium in

he myometrium or in the nutrient arteriole of submu-osa fibromyomas.11

terine CancerUterine cancer manifests as an irregular filling

efect (Fig 10), but is rarely diagnosed by the HSG

IG 7. Small size of the uterus cavity with normal length of the vagina.

IG 8. Submucosa fibromyoma. Contrast deficiency with smoothorder at the bottom of the uterus.

ethod.

203

I

epGaala

H

afitsoa

m(

T

otruw

S

uofi

F

Fb

Fo

Ff

2

ntrauterine AdhesionsIntrauterine adhesions are most commonly caused by

ndometrial trauma of curettage. They are also seen inatients with chronic endometriosis due to tuberculosis.enital tuberculosis primary affects the fallopian tubes,

nd 50% of patients with tubal disease also have a uterinebnormality.27 Intrauterine adhesions manifest as irregu-ar filling defects, most commonly as linear filling defectsrising from one of the uterine walls.4

ydrosalpinxHSG is the best method for visualizing and evalu-

ting the fallopian tubes. Hydrosalpinx is a commonnding that results from a previous inflammation of

he fallopian tubes (salpingitis). This is usually theequelae of distal tubal occlusion, leading to dilationf the proximal segment.5 The radiologic image shows

IG 9. Endometriosis. Sack-shaped projection full of contrast medium.

IG 10. Uterine cancer. Large contrast deficiency with abnormalorder at the left lateral uterus wall, which is indicated.

dilated lumen in one or more spots, and the contrast a

04

edium will not make its way to the peritoneal cavityFig 11).

uberculated SalpingitisThis entity usually causes distant fallopian tube end

bliteration. In extensive infections, multiple constric-ions along the course of fallopian tube can form,esulting in areas of dilation and stenosis.27 Abnormalterine and vaginal profiles are observed in cases ofidespread infection.

alpingitis Isthmica NodosaSalpingitis isthmica nodosa (Fig 12) is a disease of

nknown etiology, characterized by multiple smallutpouchings or diverticula affecting one or bothallopian tubes. It is presumably caused by pelvicnflammatory disease or endometriosis and is associ-

IG 11. Hydrosalpinx. Unicorn uterus (asterisk) with a dilated far endf the fallopian tube.

IG 12. Nodosa isthmic salpingitis. Presence of small projected spotsull of contrast medium, parallel to the fallopian tube.

ted with ectopic pregnancy and infertility.9

Curr Probl Diagn Radiol, September/October 2009

N

noiaceimdbwafpnn

E

iorsll

CHialwkt

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

C

ondrawing of the Fallopian TubesThis is the most common finding during the exami-

ation and is usually caused by poor technique, spasm, orbliteration of the fallopian tube. Poor technique includesmperfect straightening of the external cervical ostium orn inadequate amount of contrast medium in the uterineavity. The cornual portion of the fallopian tube isncased by the smooth muscle of the uterus and, if theres a spasm of the muscle during HSG, one or both tubesay not fill. Using radiography, tubal spasm cannot be

istinguished from tubal occlusion. This could be avoidedy progressive administration of the contrast medium or,hen the spasm occurs, administration of a spasmolytic

gent to relieve spasm, helping differentiate cornual spasmrom true occlusion.4 Obliteration is usually caused byrevious inflammation or uterine surgery and manifests asonopacification or abrupt cutoff of the fallopian tube witho free intraperitoneal spillage.

xternal AdhesionsExternal adhesions occur secondary to previous

nflammation or surgery, similar to the causes of tubalcclusion. Peritubal adhesions prevent contrast mate-ial from flowing freely around the bowel loops aseen in normal cases, and most commonly manifest asoculation of the contrast material around the ampul-ary portion of the tube.5

onclusionSG remains the front-line imaging modality in the

nvestigation of infertility. It is an accurate means ofccessing the uterine cavity and tubal patency, but has aow sensitivity for the diagnosis of pelvic adhesions,hich is why it cannot replace laparoscopy. It requiresnowledge of the female anatomy as well as skillfulechnique in order to avoid pitfalls and misinterpretations.

REFERENCES1. Baramki TA. Hysterosalpingography. Mod Trends 2005;83:1595.2. Schwabe MG, Shapiro SS, Haning RV Jr. Hysterosalpingogra-

phy with oil contrast medium enhances fertility in patients withinfertility of unknown etiology. Fertil Steril 1983;40:604-6.

3. Livsey R. Hysterosalpingography. Australas Radiol 2001;45:98-9.

4. Simpson W, Beitia LG, Mester J. Hysterosalpingography: Areemerging study. Radiographics 2006;26:419-31.

5. Eng CW, Tang PH, Ong CL. Hysterosalpingography: Currentapplications. Singapore Med J 2007;48:368-73.

6. American College of Radiology. ACR standard for the per-formance of hysterosalpingography. In: ACR Standards.

Reston, VA: ACR, 2001. p. 183-6.

urr Probl Diagn Radiol, September/October 2009

7. Soares SR, Reis MMBB, Camargos AF. Diagnostic accuracyof soundhysterosalpingography, transvaginal sonography andhysterosalpingography in patients with uterine cavity disor-ders. Fertil Steril 2000;73:406-11.

8. Lees WR, Highman SH. Gynecological imaging In: Sutton D,editor. Textbook of Radiology and Imaging. New York:Churchill Livingstone, 1988. p. 1235-72.

9. Crofton M, Jenkins PRJ. Hysterosalpingography. In: SuttonD, editor. Textbook of Radiology and Imaging, 7th edition.New York: Churchhill Livingstone, 2003. p. 1085-6.

0. Dhaliwal LK, Gupta UR, Aggarwall N. Is hysterosalpingog-raphy an important tool in modern gynecological practice? IntJ Fertil Womens Med 1999;44:212-5.

1. Meschan I. Analysis of Roentgen signs. In: Obstetrics and Gynecol-ogy, vol 3. Philadelphia, PA: WB Saunders, 1983; 1896-907.

2. Thurmond AS, Jones MK, Cohen DL, et al. Procedures for diagno-sis and treatment of infertility. In: Gynecologic, Obstetric and BreastRadiology. Cambridge: Blackwell Science, 1996. p. 114-34.

3. Steiner AZ, Meyer WR, Clark RL, et al. Oil-soluble contrastduring hysterosalpingography in women with proven tubalpotency. Obstec Gynecol 2003;101:109-13.

4. Ramsuen F, Lindequist S, Larsen C, et al. Therapeutic effects ofhysterosalpingography: oil- versus water-soluble contrast media—Arandomized prospective study. Radiology 1991;179:75-8.

5. Spring DB, Barkan HE, Pruyn SC. Potential therapeutic effectsof contrast materials in hysterosalpingography: A prospectiverandomised clinical trial. Radiology 2000;214:53-7.

6. Chalazonitis AN, Tzovara J, Tsimitselis G. Hysterosalpingog-raphy. A Pictorial Review. From: 91th Scientific Assemblyand Annual Meeting of the Radiological Society of NorthAmerica, Feb 10-12, 2005, Chicago, IL.

7. Measday B. An analysis of the complications of hysterosal-pingography. J Obstet Gynecol Br Emp 1960;67:663.

8. Siegler AM. Dangers of hysterosalpingography. Obstet Gy-necol Surg 1967;22:284.

9. Dahnert W. Obstetric and gynaecological disorders. RadiologyReview Manual. Baltimore, MD: Williams & Wilkins, 1996. p.763-4.

0. Trolano RM, McCarthy SM. Mullerian duct anomalies: Im-aging and clinical issues. Radiology 2004:233;19-34.

1. Stampe Sorensen S. Estimated prevalence of mullerian ductanomalies. Acta Obstet Gynecol Scand 1998;67:441-5.

2. Stray-Pedersen B, Stray-Pedersen S. Etiologic factors and subse-quent reproductive performance in 195 couples with a prior historyof habitual abortion. Am J Obstet Gynecol 1984:148;140-6.

3. Raga F, Bauset C, Remohl J, et al. Reproductive impact ofcongenital mullerian anomalies. Hum Reprod 1997:12;2277-81.

4. Golan A, Lnger R, Bukovsky I, et al. Congenital anomalies ofthe mullerian system. Fertil Steril 1989:51;747-55.

5. Heinonen PK, Saarikoski S, Psysynen P. Reproductive per-formance of women with uterine anomalies: An evaluation of182 cases. Acta Obstet Gynecol Scand 1983:61;157-62.

6. The American Fertility Society classifications of adnexaladhesions, distal tube obstruction, tubal occlusion secondaryto tubal ligation, tubal pregnancies, mullerian anomalies andintrauterine adhesions. Fertil Steril 1988:49:944-55.

7. Chavhan GB, Hira P, Rathod K, et al. Female genitaltuberculosis: Hysterosalpingographic appearances. BJR 2004;

77:164-9.

205