Radiol Oncol 2005; 39(1): 171-5.
Endosonography in the diagnosis of recurrent anal fistulas
Iwona Sudoł-Szopinska1, Wiesław Jakubowski1, Malgorzata Kolodziejczak2, Tomasz Szopinski3, Anna K. Panorska4
1Department of Diagnostic Imaging, Medical University Warsaw and Central Institute for
Labour Protection – National Research Institute, Warsaw, Poland; 2Subdepartment of Proctology,
Srodmiejski Hospital, Warsaw, Poland; 3Department of Urology, Central Railway Hospital,
Warsaw-Miedzylesie 4, Department of Mathematics and Statistics, University of Nevada,
Reno, USA,
Background. The aim of this work was to compare non-contrast endosonography (NCE) and contrast-en-hanced endosonography (CEE) in the diagnostics of recurrent anal fistulas.
Methods. In the years 1999-2002 we diagnosed 148 patients with anal fistulas. Fifty-one out of this group had recurrent anal fistulas, remaining had primary disease. For anal endosonography a Bruel&Kjaer scanner with 7.0 MHz transducer was used and 3% solution of hydrogen peroxide was used for CEE. In each case, NCE was followed by CEE, and results of both methods were compared. Results. The difference of percentages of correct diagnoses between NCE and CEE carried out 35.29% in a group of patients with recurrent anal fistulas (95% confidence interval 50.5% - 20.09%); while the differ-ence in a group of patients with primary anal fistulas was only 4.55% (95% confidence interval 11.09% -2.00%).
Conclusions. CEE significantly improves the efficiency of endosonography in diagnosing recurrent anal fistulas, whereas in primary fistulas the value of NCE and CEE is comparable.
Key words: rectal fistula-diagnosis; endosonography; recurrence
Introduction
The accuracy of anal endosonography (AES) in diagnosing the type of anal fistula, accord-Received 1 September 2005 Accepted 15 September 2005
Address to correspondence: Assist. Prof. Iwona Sudoł-Szopinska, MD, PhD, Zaklad Diagnostyki Obrazowej, Wojewodzki Szpital Brodnowski , ul Kondratowicza 8, 03-285 Warsaw; Fax +(48)22 326 5991; E-mail: iwsud@ciop.pl
ing to different authors, is from 25% to 100% and in cases of recurrent fistulas is the low-est.1-4 In spite of underlined difficulties re-sulting first of all from impossibilities of the differentiation of fistula with scar, this problem has still not been examined exactly. This study presents the own results of standard, non –contrast endosonography (NCE) and contrast-enhanced endosonography (CEE) in diagnosing the recurrent anal fistulas, and compares them with the ones obtained in the group of primary fistulas.
172
Sudoł-Szopinska I et al / Endosonography and recurrent anal fistulas
Figure 1a. Recurrent anterior transsphincteric fistula in a woman following obstetric anal sphincter trauma: complex fistula in NCE located in mid/low anal canal.
Methods
In years 1999-2002 AES was performed at 148 patients (86 male and 62 female, aged be-tween 15 to 73 years, average 46.3 years) with the clinical diagnosis of anal fistulas. In 51 from among 148 persons fistula had a recur-rent character. AES was performed by one ex-perienced radiologist, and patients were op-erated by surgeons from different centres, from which one drove a compact cooperation. In order to compare NCE with CEE only fis-tulas which on a day of executing research had a permeable external outlet were diag-nosed.
For AES a Bruel&Kjaer scanner 3535 with a mechanical transducer of frequency 7.0 MHz with the plastic cone or water balloon was used.2 No preparation was necessary prior to AES. Patients were examined in the left-lateral position with knees pulled up to abdomen. The study was performed in two-steps. The initial type of anal fistula using Park’s classification was defined,5 together with differentiation between simple and com-plex fistula, and the location of the internal opening of this fistula was defined. It was fol-Radiol Oncol 2005; 39(1): 171-5.
Figure 1b. Recurrent anterior transsphincteric fistula in a woman following obstetric anal sphincter trauma: simple fistula in CEE, located in the distal part of the scar tissues.
lowed by CEE, given through the external opening, and with the use of silicone catheter (Nelaton 10-fr), 1-2 ml of contrast which was 3% solution of hydrogen peroxide.3,6,7 Again one estimated the type of fistula, including the presence of extensions, and location of in-ternal opening. And then results obtained in both, NCE and CEE, were compared. For test-ing statistical differences between proportions of correct diagnoses in NCE and CEE methods for comparisons of dependent proportion were used.1 Results of NCE and CEE were compared with surgery. The interval be-tween AES and the operation did not exceed 8 days (1-8 days, average 2.8 days).
Results
In a group of 51 persons with recurrent anal fistulas one ascertained: 37 transsphincteric fistulas, 10 intersphincteric, 3 suprasphinc-teric, and 1 extrasphincteric (Table 1). Initially in NCE 40.8% of fistulas were simple, and 59.2% had extensions. After a contrast in-jection CEE showed that 82.4% of fistulas
Sudoł-Szopinska I et al / Endosonography and recurrent anal fistulas
173
Table 1. Comparison of non–contrast endosonography (NCE) and contrast-enhanced endosonography (CEE) in differentiation simple from complex fistulas in 51 patients with recurrent anal fistulas
Type and number	NCE		CEE		Surgery	
	simple	complex	simple	complex	simple	complex
transsphincters 37 intersphincteric 10 suprasphincteric 3 extrasphincteric 1	14 5 1 0	23 5 0 1	32 8 1 1	5 2 2 0	30 8 1 1	7 2 2 0
were simple, and only 17.6% were complex. From among all of complicated fistulas shown in NCE, in CEE one did not confirm the presence of extensions in 68, 9% of fistu-las; changes described as extensions repre-sented scars after the treatment of fistula (Figures 1a, 1b).
Surgery confirmed most of diagnoses of CEE and showed 78.4% of simple fistulas and 21.6% complex ones.
A statistical analysis was performed in order to qualify whether CEE is significantly more exact in differentiation simple from complex recurrent anal fistulas than NCE. In NCE correct diagnoses were obtained in 56.86%, whereas in CEE 92.16%. The differ-ence of percentages of correct diagnoses be-tween NCE and CEE carried out 35.29% and a 95% confidence interval for this difference was from 50.5% to 20.09%. A significant dif-ference was showed between NCE and CEE in differentiating of simple and complex re-current fistulas (p<0.0006), which proved that CEE is significantly more exact in differenti-ating simple from complex recurrent anal fis-tulas. The divergence between CEE and sur-gery is ascertained only in two transsphinc-teric fistulas. The injection of contrast exten-
sions were not confirmed, in spite of the fact that the former ones were visible in NCE and became confirmed during the surgery. However, the general efficiency of CEE was significantly greater than that of NCE which proved to be not reliable of differentiating scars with the active fistula. The percentage of falsely positive diagnoses of complex fistu-las in NCE carried out 36.7% (18 from 49 fis-tulas found in NCE), in CEE falsely positive results did not ascertain.
To confirm the diagnostic value of the use of contrast in the investigation of recurrent fistulas, the above results were compared with the endosonographic image of primary fistulas (Table 2). To compare this analysis more accurately, only the types of primary fistulas which were found in a group of re-current fistulas (i.e. transsphincteric, inter-sphincteric, suprasphincteric, and extras-phincteric) were included. In NCE 89.9% of simple, and 10.1% of complex fistulas were found. After the contrast administration 97.1% appeared simple and 2.99% had exten-sions. By the surgery 97% were found simple, and 3% complex, similarly as in CEE.
Identical as for recurrent fistulas a statisti-cal analysis was done for the primary fistulas.
Table 2. Comparison of non–contrast endosonography (NCE) and contrast-enhanced endosonography (CEE) in differentiation simple from complex fistulas in 66 patients with primary anal fistulas
Type and number	NCE		CEE		Surgery	
	simple	complex	simple	complex	simple	complex
transsphincteric 36 intersphincteric 13 suprasphincteric 7 extrasphincteric 10	31 12 10 9	5 1 0 1	34 13 8 10	2 0 0 0	34 13 7 10	2 0 0 0
Radiol Oncol 2005; 39(1): 171-5.
174
Sudoł-Szopinska I et al / Endosonography and recurrent anal fistulas
In NCE 93.94% of correct diagnoses were ob-tained and in CEE 98.48%. The difference be-tween percentage of correct diagnoses in NCE and CEE was 4.55% and 95% confidence interval was from 11.09% to 2.00%. A statisti-cal difference between NCE and CEE in dif-ferentiating simple from complex primary fis-tulas was found at the level of significance p=0.09. The test was not characteristic on 5% but only on 10% level. Therefore it is ascer-tained that CEE only slightly improves the di-agnostic accuracy of endosonography for the primary anal fistulas.
Discussion
Cheong et al3 underlines that CEE is especial-ly precious in diagnosing recurrent and com-plex anal fistulas. Our statistical analysis con-firmed that for recurrent anal fistulas CEE is significantly more accurate than NCE (p<0.0006). In NCE 36.7% of falsely positive diagnoses of complex fistulas were found, in CEE such results were not observed. The effi-cient treatment of fistula depends on the eradication of all extensions. However, limi-tations of AES in patients with a history of surgery of anal fistula or abscess, resulting from difficulties in differentiating scars with the active fistula, and especially with its ex-tensions, were well known.3,4,8 Although Law et al9 describes that the scar has lower and more homogeneous echogenicity than the fistula and smooth outlines as well; the most of-ten image of these two is identical. Additi-onally narrow, irregular lumen of the recur-rent fistula and its extensions has often no content liquid or air, which have a character-istic image.6 Consequently not recognized and not removed extensions are the main reasons of the recurrence of fistula, and a wrong estimated type of fistula can lead to damages of anal sphincters.3,4,8
In spite of underlined difficulties with en-dosonographic diagnostics of recurrent fistu-Radiol Oncol 2005; 39(1): 171-5.
las, one did not examine the scale of this problem. This study confirmed a large num-ber of false diagnoses of complex fistulas in NCE. A comparative analysis with primary fistulas showed that in a group of primary fis-tulas it had a place only in 7.2% of fistulas. In NCE scars following previous surgery were interpreted as extensions, and so the accura-cy in such differentiation was only 56.86%. The introduction of contrast raised to 92.16%. Also Cheong et al3 and Kruskal et al6 empha-sized that only CEE can be accurate. Using CEE Kruskal et al accurately differentiated scars with fistulas in 20 from 30 patients (67%), including 39 patients with a doubtfully initial, without contrast, image. Our results showed that in only in two cases (4%) exten-sions of transsphincteric fistulas were not recognized in CEE. One ran in the direction of the top of ischio-rectal fossa, the second crossed the levator ani muscle. The former was probably blocked by thick secretion, and the latter would become visible when AES was supplemented by the use of a water bal-loon. In this case, however, too hastily initial NCE became interpreted as scar and one did not extend the investigation of ampulla of the rectum. It seems that such an approach should be done in case of high fistulas, espe-cially the recurrent ones. The other thing is that, in spite of the proved significantly higher accuracy of CEE, one should also taken in-to account the result of NCE. Results ob-tained in a group of primary fistulas (66 pa-tients) diametrically differed from those of the recurrent ones.
The number of complex fistulas was not large, both in NCE and in CEE, and the per-centage of falsely diagnosed complex fistulas in NCE was only 7.24%. As in a case of recur-rent fistulas, the scars after the treatment of fistula were responsible for correct diag-noses; so, in the primary fistulas the only rea-son was the inability to differentiate exten-sions from the heterogeneous echotexture of perirectal tissues. A statistical analysis showed
Sudoł-Szopinska I et al / Endosonography and recurrent anal fistulas
175
that NCE has comparable values to CEE. However, although it seems that it is more important not to miss the extension that gives a false diagnosis of extension (regarding the risk of surgical complications: recurrence, damage of sphincters), one must remember that too aggressive approach during the operation, in order to find indicated in NCE ex-tensions, can also lead to complications - cre-ating of iatrogenic fistula.
8.   Halligan S. Review imaging fistula-in-ano. Clin Radiol 1998; 53: 85-95.
9.   Law PJ, Talbot RW, Bartram CI, Cuesta MA, Meuwissen SGM. Anal endosonography in the evaluation of perianal sepsis and fistula in ano. Br J Surg 1989; 76: 752-5.
Conclusions
1. Standard, NCE is not reliable method in differentiation scars with active recurrent fistula and application of contrast significantly improves efficacy of AES.
2. In the case of primary anal fistulas, NCE and CEE have comparable efficiency.
References
1.   Agresti A. Categorical data analysis. 2nd edition. New Jersey: John Wiley & Sons; 2002.
2.   Bartram CI, Frudinger A. Handbook of anal en-dosonography. Petersfield: Wrightson Biomedical Publishig LTD; 1997.
3.   Cheong DMO, Nogueras JJ, Wexner SD, Jagelman DG. Anal endosonography for recurrent anal fis-tulas: image enhancement with hydrogen peroxide. Dis Colon Rectum 1993; 36: 1158-60.
4.   Choen S, Nicholls RJ. Anal fistula. Br J Surg 1992; 79: 197-205.
5.   Parks AG, Gordon PH, Hardcastle JD. A classifi-cation of fistula-in-ano. Br J Surg 1976; 63: 1-12.
6.   Kruskal JB, Kane RA, Morrin MM. Peroxide-en-hanced anal endosonography: technique, image interpretation, and clinical applications. Radiographic 2001; 21: 51-73.
7.   Poen AC, Felt-Bersma RJF, Eijsbouts QAJ, Cuesta MA, Meuwissen SGM. Hydrogen peroxide-en-hanced transanal ultrasound in the assessment of fistula-in-ano. Dis Colon Rectum 1998; 41: 1147-52.
Radiol Oncol 2005; 39(3): 171-5.