Introduction

The anatomic sites of occurrence of pelvic endometriosis have been represented by pelvic diagrams [1, 2] and tabulations [3, 4, 5, 6, 7, 8]. Such diagrams and tabulations are cumulative frequency counts, and older publications may have selection bias if subtle or atypical forms of endometriosis were not noted accurately.

Computer tabulation of each anatomic area of involvement in each patient would result in a more dynamic mapping system that might aid in understanding the disease. Accurate identification of subtle as well as typical lesions would increase the accuracy of this system. This report is the first to describe this computerized mapping system and preliminary results of its study with respect to the extent of disease in the presence or absence of ovarian endometriosis.

Materials and methods

The Endometriosis Treatment Program of St. Charles Medical Center in Bend, Oregon, is a tertiary center for surgical treatment of endometriosis-associated pain. Patients are treated by aggressive excision of visually abnormal peritoneum or serosa, including resections of the diaphragm, urinary, and gastrointestinal tracts. Excision might be superficial or deep, depending on the degree of invasive fibrosis associated with the disease. The result of excision is removal of all abnormal visual changes and invasive fibrosis so that only healthy tissue remains.

Intestinal endometriosis is treated by partial-thickness, full-thickness, or segmental bowel resection as required by the individual case for complete excision. Excised tissue is placed into bottles labeled by anatomic pelvic regions for histologic study, resulting in >40 individual biopsy specimens in some cases. All surgery was performed by the author, with a general surgeon available for segmental bowel resections. The presence of glands and stroma on microscopy was required to confirm endometriosis for entry into the computer database.

Individual sites of biopsy-proven pelvic or intestinal disease for each patient are entered into a computer database application (FoxPro for the Macintosh; Microsoft, Inc., Redmond, WA). From 1979 to 1988, sites of pelvic involvement were hand tabulated on accounting ledgers, and study of that database has been the subject of previous work [9, 10]. Since 1988, sites have been computer tabulated, and the previous manually acquired data were entered also. All data are entered contemporaneously by the author, usually within a few days of surgery after the operative report, final pathology report, and a surgical diagram of findings have been assembled to ensure accurate data entry. The database contains 1,979 patients with endometriosis, among whom 547 had intestinal involvement. The most recent date of surgery in this series was January 28, 1999.

Pelvic mapping of individual sites of pelvic or intestinal endometriosis gives an accurate representation of the geographic extent of anatomic involvement by disease. The revised classification system of endometriosis of the American Society of Reproductive Medicine [11] cannot be used to accurately gauge pelvic involvement, because most of the available points score for adhesions, not endometriosis. In addition, in that system, most endometriosis points are assigned to the ovaries, which are not the most common site of pelvic involvement. Finally, the extent of peritoneal and intestinal involvement cannot be discriminated accurately because a maximum of six points is allowed for peritoneal involvement anywhere in the pelvis, whereas intestinal disease is not scored in that classification system.

Management of this computer database has been outlined elsewhere [12, 13, 14]. Although the database currently contains 139 fields, for purposes of this study the areas of pelvic and intestinal involvement and previous therapy will be considered. Five separate intestinal areas and 15 separate pelvic areas are included in the database. No patient had more than 12 pelvic areas of involvement.

Because each ovary has a separate logical field in the database to indicate involvement by endometriosis, in each patient with ovarian disease the number of pelvic areas of involvement was reduced by the number of ovaries involved by endometriosis in that patient. This allows direct comparison of nonovarian areas of pelvic involvement in the group with ovarian disease to the group without ovarian disease. One hundred ninety-four patients with endometriosis with previous hysterectomy or hysterectomy with castration were excluded because the uterus and ovaries are separate anatomic sites listed in the database. The database of the remaining 1,785 patients is analyzed with programs written by the author in FoxPro source code.

To evaluate whether large ovarian endometriomas versus superficial ovarian endometriosis had any effect on findings, the data were stratified around a revised American Fertility Society (AFS) endometriosis-only point score of 11. Patients with fewer than 11 endometriosis points would include those with only superficial ovarian disease or small endometrioma cysts of <1 cm in diameter.

Previous medical or surgical therapies for endometriosis were also tabulated in the database. Because previous medical or surgical therapy might be expected to reduce the number of pelvic areas involved by disease, the effect of previous therapy also was analyzed to determine whether it is important to correct such an analysis for previous therapy. Previous surgical therapies usually included thermal ablation techniques, such as endocoagulation, laser vaporization, or electrocoagulation, but infrequently included removal of ovarian endometriomas or small peritoneal biopsies. Previous medical therapies included GnRH agonists, danazol, or oral contraceptive steroids usually given cyclically, or rarely, continuously.

Not all medical and surgical therapies had necessarily been used in each previously treated patient. A substantial minority of patients (450 [25%]) had no previous therapy. Patients in this series with intestinal endometriosis rarely had a previous correct diagnosis of intestinal involvement and almost never had previous surgical treatment of intestinal disease before surgery by the author, so previous therapy would not be expected to affect results of the analysis of intestinal involvement.

Mean number of pelvic or intestinal sites of involvement by endometriosis with confidence intervals and risk ratios with confidence intervals were used to analyze the data.

The pelvic mapping system described in this study had no effect on the surgical management of any patient and is an expansion of the mapping diagram supplied by the AFS in its revised classification system [11]. This mapping diagram is in widespread use around the world as a recommended part of accepted clinical practice in the care of patients with endometriosis. Because of these considerations, institutional review board review of this study was not requested.

Results

The frequency of involvement of pelvic and intestinal sites by endometriosis in all study patients is tabulated in Table 1. The ovaries are not the most common pelvic site of occurrence of endometriosis.

Note: Total of percentages exceeds 100% because most patients had more than one area of involvement.
*Four patients had only intestinal endometriosis with no pelvic involvement.
†One hundred and ninety-four patients with endometriosis with previous hysterectomy with or without castration have been excluded.
‡One hundred and fifty-four patients had more than one intestinal area of endometriosis.
§Some patients had previous appendectomy.

Not all peritoneal biopsies returned as endometriosis (Table 2). These biopsies were not included as endometriosis in the database.

Note: Some patients had more than one nonendometriotic finding. Nonendometriotic findings were not tabulated as endometriosis.

Previous therapy did not alter the finding of more extensive pelvic and intestinal disease when ovaries were involved (Table 3).

Previous therapy did not alter the finding of more extensive pelvic and intestinal disease when ovaries were involved (Table 3).

Compared with deep ovarian endometriosis manifested as endometrioma cysts, superficial ovarian endometriosis was associated with slightly more extensive pelvic disease but less extensive intestinal disease (Table 4).

Ovarian endometriosis was associated with an increased risk of intestinal involvement, and this trend was not affected by previous therapy (Table 5). When ovarian endometriosis and intestinal endometriosis coexisted, the risk of full-thickness or segmental bowel resection for complete removal of disease was increased (Table 6).

Among 1,785 patients, only 19 (1.06%) had ovarian involvement exclusively with no other pelvic areas of involvement.

Discussion

Since Cullen's classic description of adenomyomas of the rectovaginal septum and other pelvic sites [1], pelvic diagrams have played an important role in educating physicians about where endometriosis can be located. Most often, these diagrams have been simply frequency counts of where endometriosis can be located in the pelvis in a relatively small series of patients.

The accuracy of such diagrams can be degraded by lack of biopsy control and dependence only on the visual appearance of endometriosis, because not all pelvic peritoneal abnormalities are endometriosis (Table 2). The accuracy of these historical diagrams can also be affected by the tabulating surgeon's lack of awareness of the protean visual manifestations of endometriosis. If surgeons identify only black "powderburn" lesions but miss more subtle disease, then pelvic diagrams suffer from selection bias and may transmit erroneous information.

Peritoneal disease is more common than ovarian disease (Table 1), contradicting the previously long-held clinical dictum that the ovaries are the most common sites of involvement of endometriosis [15, 16, 17, 18, 19, 20, 21]. The notion that the ovaries are the most common location of pelvic endometriosis probably resulted from the fact that ovarian endometriomas are usually quite obvious at surgery, but subtle or even deeply invasive disease hidden beneath might go unnoticed because the main surgical thrust frequently would be simply to treat the ovarian masses, not to perform the difficult pelvic dissection that would allow the diagnosis and treatment of hidden or invasive disease. Even Sampson eventually realized that peritoneal involvement was the most frequent and clinically the most important form of endometriosis [22].

When corrected for the occurrence of their ovarian disease, patients with ovarian endometriosis have more pelvic and intestinal areas involved than patients without ovarian involvement. A history of previous medical or surgical therapy had no effect on this pattern. In fact, patients with previous therapy had more extensive pelvic and intestinal involvement than patients without previous therapy. Medical therapy rarely, if ever, eradicates even minimal endometriosis [23] and, therefore, would not be expected to have any effect on extent of the disease.

There are three possible explanations for this trend of more extensive disease in previously treated patients. First, patients attracted to a referral center may have more extensive involvement than patients seen in local practice, and previous surgical therapy may have reduced it only slightly, resulting in a persistent prevalence of more extensive disease. Second, patients with more extensive disease and undergoing previous surgical therapy may not have had any effect on the extent of their disease at all. Third, previous medical or surgical therapy may have promoted more extensive disease, although this seems unlikely.

Extent of ovarian endometriosis was not correlated with extent of pelvic disease. Compared with superficial ovarian endometriosis, deep ovarian disease (endometrioma cysts >1 cm in diameter) was associated with slightly less extensive pelvic disease but more extensive intestinal disease. Thus, there is no suggestion that as the ovary "fills up" with disease that so does the pelvis. It appears that the occurrence of any ovarian disease is accompanied by an immediate increased likelihood of more extensive pelvic disease. This may speak to a process of origin of disease shared by the pelvic mesothelium and by the ovarian germinal epithelium.

Of 482 patients with ovarian endometriosis, 105 (21.7%) had complete cul-de-sac obliteration and 9 (2%) had partial obliteration. Of 1,303 patients without ovarian endometriosis, 75 (5.8%) had complete cul-de-sac obliteration, and 16 (1.2%) had partial obliteration. Among patients with partial obliteration of the cul-de-sac in this study, the rectal wall was involved by endometriosis in 60% of cases, whereas the rate of rectal wall involvement was 70% in patients with complete cul-de-sac obliteration.

According to the results of this study, patients with ovarian disease and intestinal involvement are more likely to require full-thickness or segmental bowel resection for complete removal of their disease compared with patients with intestinal involvement but without ovarian endometriosis. Therefore, patients with known endometriomas or with imaging tests suggesting the presence of endometriomas should be bowel prepped before surgery. Patients with known obliteration of the cul-de-sac or with significant pelvic nodularity should also receive preoperative bowel preparation. A general surgeon should be on call if the surgical team expects to be able to removal all intestinal disease.

In this study, the ovaries rarely were exclusively involved by endometriosis without coexistent disease elsewhere. This indicates that if a surgeon identifies and treats only ovarian endometriosis, there is a 98.9% likelihood that other pelvic or intestinal disease is being left untreated. This may contribute to the repetitive surgeries that are the unfortunate hallmark of modern therapy for endometriosis.

The current study does not suffer from small numbers or lack of awareness of subtle forms of endometriosis. All patients were operated on by a surgeon with demonstrated awareness of known manifestations of pelvic and intestinal endometriosis, including so-called "microscopic" disease. All disease was treated by aggressive excision, resulting in biopsy proof of endometriosis. Nonendometriotic lesions (Table 2) were not included in the database as endometriosis.

Possible confounding conditions exist. Previous medical and surgical therapies were tabulated on the basis of a review of all available medical records and the patient's oral and written history. Sometimes, not all records were available for review, and patient recall is not always accurate. In the current study, there is no way to detect or correct reporting errors. All data entry was done by the author. In any large database with hundreds of thousands of entries, errors in data registration are possible. However, during routine retrieval of patient administrative data (names, addresses, etc.) errors have been found infrequently, and it is unlikely that entry of clinical data would be more susceptible to errors.

The referral patients in this study may differ in some unknown way from patients seen by local practitioners. This possibility prompted the examination of patients with no previous therapy, because these patients might be more representative of nonreferral patients. Nonetheless, patients without previous therapy exhibited the same findings as previously treated patients. It seems likely that these findings also would apply to patients seen by local practitioners.

Complete removal of endometriosis was the surgical goal in each patient. Excision differs fundamentally from thermal ablation techniques commonly used to treat endometriosis, and thermal ablation techniques would be unlikely to produce findings similar to those in this study relating to intestinal disease. For example, excision allows full-thickness or segmental bowel resection as needed for identification and removal of invasive intestinal endometriosis. Surgeons using thermal ablation techniques would seek to avoid penetration into the wall of the bowel, and invasive intestinal disease would thus be "converted" into superficial disease because of the limitations of thermal ablation techniques.

These results have important practical points for clinicians caring for patients with endometriosis. Ovarian endometriosis appears to be a marker for more extensive pelvic and intestinal disease compared with patients without ovarian endometriosis. Almost 100% of patients with ovarian disease have disease elsewhere in the pelvis or intestinal tract. Surgeons identifying and treating only ovarian endometriosis may be underdiagnosing and undertreating their patients. Since ovarian endometriosis is associated with an increased frequency of invasive intestinal disease, consideration should be given to preoperative bowel preparation and the assistance of a general surgeon. Partial or complete obliteration of the cul-de-sac usually requires some degree of intestinal surgery for complete removal of endometriosis.

References

1. Cullen TS. The distribution of adenomyomas containing uterine mucosa. Arch Surg 1920;1:215-83.
2. Fallon J. Endometriosis: evidence of tubal origin in the distribution of lesions. Arch Surg 1951;62:412-9.
3. Henrikson E. Endometriosis. Am J Surg 1955;90:331-7.
4. Smith GVS. Endometrioma: a clinical and pathologic study of 159 cases treated at the clinic of the free hospital for women, Brookline, Massachusetts. Am J Obstet Gynecol 1929;17:806-14.
5. Pemberton FA. Endometrioma of the female genital organs. N Engl J Med 1937;217:1-5.
6. Counseller BS. Endometriosis: a clinical and surgical review. Am J Obstet Gynecol 1938;36:877-88.
7. Fallas R, Rosenblum G. Endometriosis: a study of 260 private hospital cases. Am J Obstet Gynecol 1940;39:964-75.
8. Holmes WR. Endometriosis. Am J Obstet Gynecol 1942;43:255-66.
9. Redwine DB. The distribution of endometriosis in the pelvis by age groups and fertility. Fertil Steril 1987;47:173-5.
10. Redwine DB. Age related evolution in color appearance of endometriosis. Fertil Steril 1987;48:1062-3.
11. The American Fertility Society. Revised American Fertility Society classification of endometriosis: 1985. Fertil Steril 1985;43:351-2.
12. Redwine DB. Database management for personal computers, I. J Am Assoc Gynecol Laparosc 1996;3:469-74.
13. Redwine DB. Database management for personal computers, II. J Am Assoc Gynecol Laparosc 1996;3:643-50.
14. Redwine DB. Database management for personal computers, III. J Am Assoc Gynecol Laparosc 1996;4:95-101.
15. Cattell RB, Swinton NW. Endometriosis: with particular reference to conservative treatment. N Engl J Med 1936;214:341-6.
16. Doughert CM, Anderson MR. Endometriosis and adenomyosis. Am J Obstet Gynecol 1964;89:23-37.
17. Acosta AA, Buttram VC, Besch PK, Malinak LR, Franklin RR, Vanderheyden JD. A proposed classification of pelvic endometriosis. Obstet Gynecol 1973;42:19-24.
18. Wild RA, Wilson EA. Clinical presentation and diagnosis. In: Wilson EA, ed. Endometriosis. New York: Alan R Liss, 1987:53-77.
19. Jenkins S, Olive DL, Haney AF. Endometriosis: pathogenic implications of the anatomic distribution. Obstet Gynecol 1991;67:335-8.
20. Kruitwagen RFPM, Poels LG, Willemsen WNP, Jap PHK, Ronde IJY, Hanselaar TGJM, et al. Immunocytochemical markerprofile of endometriotic epithelial, endometrial epithelial, and mesothelial cells: a comparative study. Eur J Obstet Reprod Biol 1991;41:215-23.
21. Schwartzwald D, Mooppan UMM, Ohm HK, KimH. Endometriosis of bladder. Urology 1992;39:219-22.
22. Sampson JA. The development of the implantation theory for the origin of peritoneal endometriosis. Am J Obstet Gynecol 1940;40:549-57.
23. Redwine DB. Treatment of endometriosis-associated pain. In: Olive DL, ed. Endometriosis: Infertility and Reproductive Medicine Clinics of North America. Philadelphia: WB Saunders 1992:697-720.