Is cystic ovarian endometriosis an asymmetric disease?

Authors


Correspondence: Dr P. Vercellini, Clinica Ostetrica e Ginecologica “Luigi Mangiagalli”, Università di Milano, Via Cornmenda, 12, 20122 Milano, Italy.

Abstract

Objective To investigate whether asymmetry exists in the left- and right-handed distribution of ovarian cystic lesions in a large series of women with endometriosis.

Design Retrospective evaluation of a case series.

Setting Tertiary care and referral academic centre for the study and treatment of endometriosis.

Population A total of 1054 consecutive women undergoing first-line surgical treatment for endometriosis in an eight-year period.

Methods Data were collected on indication for the intervention, age at surgery, parity and disease stage as well as side and size of ovarian endometriomas.

Main outcome measure Frequency of left- and right-sided ovarian endometriomas.

Results Histologically confirmed endometriotic ovarian cysts were present in 561 women, which were on the left side in 255 instances, on the right in 148, and bilateral in 158. In the patients with unilateral endometriomas, the observed proportion of left cysts (255/403, 63%; 95% confidence interval, 58% to 68%) was significantly different from the expected proportion of 50% (χ21, 28.41, P < 0.001). Including also the bilateral endometriotic cysts gave a total of 413/719 (57%) left-sided and 306/719 right-sided endometriomas. The magnitude of these proportions did not vary appreciably during the eight years considered. The difference in proportion of left- and right-sided endometriotic cysts was virtually similar in subgroups of women with different indications for surgery. Cyst side was not related to age, parity or cyst diameter.

Conclusions The finding of a lateral asymmetry in the occurrence of ovarian endometriotic cysts is compatible with the anatomical differences of the left and right hemipelvis and supports the menstrual reflux theory.

INTRODUCTION

The pathogenesis of ovarian endometriotic cysts is controversial. An endometrioma may be the result of either metaplasia of the coelomic epithelium covering the ovary1 or inversion and progressive invagination of the ovarian cortex after adhesion to the pelvic peritoneum caused by local implantation of endometrium regurgitated through the tubes2,3. Investigating the anatomical distribution of endometriotic lesions may provide insights into the pathogenesis of the disease4. According to Jenkins et at.4 if retrograde menstruation is the source of ectopic endometrium, the pattern of lesions should be determined mainly by gravity, proximity to the site of abdominal entry and anatomicophysiological variables, whereas if coelomic metaplasia is the cause of endometriosis, lesions should not be distributed in relation to factors influencing the spreading and implantation of endometrial cells in the pelvis. It has been demonstrated that sites of peritoneal implants are compatible with the former concept.4 Asymmetry in location of peritoneal lesions has been studied in terms of frequency of implants in the anterior compared with the posterior cul-de-sac, i.e. according to a sagittal view of the pelvis.4,5. The evidence available on ovarian endometriosis is limited and not consistent.4,6. Given this background we decided to investigate if an asymmetry exists also in the distribution of left- and right-sided ovarian cystic lesions in a large series of women with endometriosis.

METHODS

Clinical records were retrieved of consecutive women with endometriosis undergoing first-line conservative or definitive surgery between January 1990 and December 1997 at the First Department of Obstetrics and Gynaecology of the University of Milan. Women with genital malformations were excluded as were those who had undergone previous abdominal surgery except appendectomy. Data were collected on indication for and age at surgery, parity, disease stage according to the revised American Fertility Society classification7, and side and size of ovarian endometriomas. When two or more endometriotic cysts were present in the same gonad, only the one with the largest diameter was considered. Ovarian endometrioma was diagnosed by the pathologist in the presence of two or more of the following microscopic patterns: endometrial epithelium, endometrial glands or gland-like structures, endometrial stroma, and haemosiderin-laden macrophages8.

The frequency of left- and right-sided cystic ovarian endometriosis in the entire series was analysed with the χ2 test to compare observed and expected events. The confidence interval (CI) of the proportion of left endometriomas was computed using the normal approximation. The odds ratio (OR) of an endometrioma occurring on the left compared with the right ovary was computed separately for various factors from data stratified for quinquennia of age using logistic regression9. 95% CI for OR were based on profile likelihood. When a factor could be classified in more than two levels, the significance of the linear trend was assessed by the Mantel test for nonzero correlation, otherwise heterogeneity was tested by the Cochrane-Mantel-Haenszel test for general association9.

RESULTS

In the study period 1054 women underwent first-line surgical treatment for endometriosis at laparoscopy or laparotomy. The vast majority of them were women self-referring for various conditions to a tertiary care academic centre for the treatment of endometriosis. The disease was at stage I in 319 (30%) cases, Stage II in 138 (13%), Stage III in 293 (28%), and Stage IV in 304 (29%). A total of 561 women had histologically confirmed endometriotic ovarian cysts which were on the left side in 255 instances, on the right in 148, and bilateral in 158. In the patients with unilateral endometriomas the observed proportion of left cysts (255/403, 63%, 95%, CI, 58% to 68%) was significantly different from the expected proportion of 50% (χ21 28.41, P < 0.001). Including also the bilateral endometriotic cysts gave a total of 413/719 (57%) left-sided and 306/719 right-sided endometriomas. The magnitude of these proportions did not vary appreciably during the eight years considered. Table 1 shows the distribution of left and right unilateral ovarian endometriotic cysts according to selected factors. The percentages of left and right unilateral endometriomas were not significantly different in the subgroups of women with various indications for the intervention and were consistently unbalanced in favour of the left side. In 34 cases the diagnosis was incidental at the time of procedures such as myomectomy or hysterectomy. Exclusion from the analysis of the women with an incidental diagnosis who had conditions not related to specific symptoms of endometriosis did not modify the overall result (data not shown). No significant association emerged between age at surgery, parity and proportion of left-versus right-sided endometrioma. Likewise, we observed no relation between cyst diameter and side. The above results were generally similar when also the bilateral cysts were considered.

Table 1.  Frequency of left- and right-sided unilateral ovarian endometriomas according to indication for and age at surgery, parity, and cyst diameter (Milan, Italy: 1990–1997).
 Left endometriomaRight endometrioma 
 n(%)n(%)OR (95%CI)
  1. aAdjusted for age in quinquennia.

  2. bReference category.

  3. χ23 =χ2 test with 3 degrees of freedom.

  4. χ212 test with 1 degree of freedom.

Indication for surgery
  Infertility95(66)50(34) 
  Pelvic pain61(65)33(35)χ23a for heterogeneity
  Adnexal mass85(61)55(39)= 0.55;P= 0.91
  Other14(58)10(42) 
Age at surgery (years)
  <2514(5)6(4)1b
  25–2963(25)29(20)0.93 (0.30–2.58)
  30–3489(35)55(37)0.69(0.23–1.84)
  35–3963(25)32(22)0.84(0.28–2.32)
  >4026(10)26(17)0.43(0.13–1.25)
  χ21 trend    3.13; P= 0.08
Parity
  0190(75)110(74)1b
  147(18)20(14)1.51 (0.85–2.78)
  ≥218(7)18(12)0.74(0.34–1.60)
  χ21 trenda    0.14; P= 0.71
Cyst diameter (cm)
  <17(3)7(5)1b
  1–341(16)24(16)1.64(0.50–5.37)
  >3–595(37)47(32)1.87(0.60–5.82)
  >5112(44)70(47)1.46(0.48–4.47)
  χ21 trenda    0.00; P= 0.99

DISCUSSION

The ovary is the most common site of endometriotic lesions.4,6 Jenkins et al. have already reported a more frequent involvement of the left (81/182, 4%), compared with the right (57/182, 31%) gonad in a consecutive series of women with laparoscopically diagnosed endometriosis. The results of our study of a larger series confirm that ovarian endometriomas are found more frequently on the left than the right side. The difference in proportion of left- and right-sided endometriotic cysts was virtually similar when we considered subgroups according to parity and indications for surgery, which supports the consistency of the general results. A trend was observed towards an association with age (P= 0.08), but this was mainly due to the distribution of the relatively small number of cysts found in women of over 40, and random fluctuation of data cannot be excluded.

Information was abstracted from clinical records by a single experienced research fellow. The accuracy of the process was not formally evaluated. Specifically, we did not match data abstracted by more than one independent reviewer. However, all the diagnoses were based on the pathological report and there were no exclusions apart from those specified in the study protocol. Moreover, the surgeons who recorded the clinical data, the pathologists who performed the histological examination and the reviewer were all unaware of the hypothesis of the study. The surgical approach was mixed but it is improbable that pelvic visualisation at laparoscopy versus laparotomy could have influenced the detection of ovarian endometriomas8,10, as it would in the case of limited peritoneal lesions. Furthermore, most of the surgical procedures were performed under the supervision of a senior researcher. Finally, in the context of a teaching programme, all our patients undergo pelvic ultrasound examination before surgery. This should have revealed also small cysts not readily identifiable at visual inspection.11

According to the proponents of the metaplasia theory, endometriotic cysts derive from invagination and differentiation of coelomic epithelium into the ovarian cortex1. This process, based on the metaplastic potential of pelvic mesothelium, is closely linked to the pathogenetic mechanism of epithelial ovarian tumours. However, if this is the case, no major asymmetry in the frequency distribution of left- and right-sided ovarian endometriomas should be expected. Furthermore, no such asymmetry has been definitely demonstrated for epithelial ovarian cancers12, although it has been reported that the right ovary may be more frequently affected13. The same considerations apply to the theory suggesting that large endometriomas may develop as a result of secondary invasion of luteal cysts by ovarian surface endometriotic cells14. In fact, according to the available evidence, ovulation occurs either equally on both sides15 or more frequently on the right16 but not the left side. If instead retrograde menstruation is the origin of endometriotic lesions2,3,17 our findings should be considered taking into account anatomic factors that may influence the distribution and implantation of regurgitated endometrial cells.4,18,19 Indeed, the two adnexal regions are different in terms of exposure to the pelvic milieu, the left one being ‘protected’ by the sigmoid colon, as is well known to those who perform gynaecologic laparoscopic surgery. Not only does this portion of the large bowel lean on the left tube and ovary but it is very often fixed to the pelvic brim by filmy adhesions which are so frequently observed as to be considered a para-physiological finding. A microenvironment is established around the left adnexa, the boundaries of which are the pelvic side wall, the lateral aspect of the sigmoid and the left broad ligament. As a consequence, endometrial cells regurgitated through the left tube are not exposed to the clockwise peritoneal current that keeps the peritoneal fluid circulating20–22 and may be partly protected from the macrophage disposal system18,19. These factors may facilitate adhesion, implantation, and growth of endometrial cells. Furthermore, involvement of the sigmoid in endometriotic lesions of the left ovary is much more common than that of the caecum, which is anatomically more cranial, in endometriomas of the right gonad. The frequency distribution of endometriotic cysts according to different diameters was largely similar for both ovaries, suggesting that, once established, the growth of an endometrioma is independent of anatomical variables.

In conclusion, our study demonstrated a lateral asymmetry in the location of ovarian endometriotic cysts. This is compatible with the anatomical differences of the left and right hemipelvis, is in line with epidemiological data suggesting that the major pathogenetic mechanism of peritoneal and ovarian endometriosis is not substantially different23, and supports the menstrual reflux theory.

Ancillary