Correspondence: Dr P. Vercellini, Clinica Ostetrica e Gineco-logica ‘Luigi Mangiagalli’, Universita di Milano, Via Commenda 12, 20122 Milano Italy.
Six cases of endometriosis obstructing the left ureter were observed among 1054 consecutive patients undergoing surgery in an eight-year period. In addition, 125 women with ureteral endometriosis (left-sided, n= 66; right-sided, n= 40; bilateral, n= 19) were described in 62 articles identified in a systematic review of the English language literature between 1980 and 1998. Considering only the patients with unilateral ureteral endometriosis and combining the published figures with those of our surgical series, the observed proportion of left lesions (72/112, 64%; 95% CI 55% to 73%) was significantly different from the expected proportion of 50% (χ2i, 9.14, P= 0.002). The lateral asymmetry found in the location of ureteral endometriosis is compatible with the menstrual reflux theory and with the anatomical differences of the left and right hemipelvis
The aetiology of endometriosis is not completely understood, and controversies exist on the pathogenesis of different forms. Ureteral lesions are relatively rare but may cause major morbidity as silent loss of renal function is not infrequent in these patients. According to Stanley et al.1 endometriosis of the ureter usually arises by extension from pelvic foci, and ovarian endometriosis is a prerequisite for ureteral involvement.
Recently we have observed that endometriotic cysts affect the left ovary more frequently than the right one and suggested that this finding is compatible with the menstrual reflux theory and with the anatomical differences of the left and right hemipelvis2,3. In fact, the presence of the sigmoid colon creates a hidden microenvironment around the left adnexa so that endometrial cells regurgitated through the left tube are less exposed to the peritoneal fluid current and may be partly protected from the macrophage disposal system. The large bowel does not provide the right hemipelvis with this sort of anatomical shelter since the caecum is more cranial2,3
The ureter is in anatomic contiguity with the lateral gonadal aspect, which has been indicated as the site of origin of the classic chocolate cyst2. Consequently, if ureteral endometriosis develops from ovarian implants or if both lesions have a common pathogenesis, asymmetry should be found also in the left- and right-handed distribution of ureteral foci.
It has been observed that left ureteral lesions are more common than right ones4, but the evidence in this regard is vague, and the issue has never been approached formally. To verify this hypothesis we reviewed the data of a large series of women with endometriosis. However, because endometriosis of the ureter is not frequent, we decided to combine our findings with those published on the topic in the scientific literature since 1980
Clinical records were retrieved for 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 and those who had undergone previous abdominal surgery, except appendectomy, were excluded. Indications and age at surgery, parity, disease stage according to the revised American Fertility Society classification5, and site and side of ureteral lesions were recorded. Ureteral endometriosis was defined as disease involvement causing luminal stenosis or obstruction with secondary hydroureter and hydronephrosis. Cases with nonobstructing lesions or forms with simple ureteral kinking or dislocation were excluded.
Moreover, with the aim of identifying all English language medical papers published on ureteral endometriosis, we conducted a MEDLINE and EMBASE search from 1980 to 1998, using the medical subject heading terms endometriosis, ureter and urinary tract. Additional reports were collected by systematically reviewing all references from retrieved papers and by consulting gynaecological and surgical textbooks and monographs on endometriosis published in the last 12 years. We considered only articles in which it was possible to assess clearly the presence of a partially or completely obstructing lesion as well as the affected side. Proceedings of scientific meetings were not included. Two authors (P.V. and A.P.) abstracted data in an unblinded fashion on standardised forms. An initial screening of the title and abstract of all articles was performed to exclude citations deemed irrelevant by both observers. The year of publication, clinical characteristics of subjects, results of pre-operative diagnostic investigations, and surgical details were recorded independently. The number of women with bilateral ureteral endometriosis and the side of those with unilateral lesions were obtained from individual studies.
The combined frequency of left- and right-sided ureteral endometriosis in the clinical series from our institution and in published reports was analysed with the χ2 test to compare observed and expected events. The 95% confidence interval of the proportion of endometriosis of the left ureter was computed using the normal approximation.
In the study period 1054 women underwent first-line surgical treatment for endometriosis at laparoscopy or laparotomy. Baseline clinical characteristics of these women have already been described2. Six women (0.6%) had obstructing ureteral endometriotic lesions which were all on the left side and associated with hydroureter, as shown by intravenous urography or retrograde ureteropyelography (Fig. 1). The stricture was always confined to the distal third of the organ, being at the level of the ovarian fossa in three instances and in correspondence with the uterosacral ligament in three. Endometriosis was at Stage IV in four cases and at Stage III in two. All the women had a left ovarian endometri-oma and four rectovaginal lesions with infiltration and fibrosis of the left parametrium. No case of bilateral involvement or association with bladder detrusor endometriosis was observed. Ureteroneocystostomy with submucosal tunnel and bladder psoas hitching was performed in two patients, ureterolysis with dissection of the ureter from the involved pelvic peritoneum through the parametrium in two (in one spiral titanium stents were also positioned), excision of the stricture with ureteroureterostomy in one, and a nephroureterec-tomy in one case with a nonfunctioning kidney.
Seventy-three articles were identified in the literature search. Of these, we excluded 11 because it was not possible to identify the lesion side or because no original data were reported. The mean number of cases observed was two (range 1-8), and the mean age (SD) was 39 years (11) (range 21-74). A total of 125 women had endometriosis of the ureter, which was left-sided in 66 instances, right-sided in 40, and bilateral in 19. Considering only those with unilateral ureteral endometriosis and combining the above figures with those of our surgical series, the observed proportion of left-sided lesions (72/112, 64%; 95% CI, 55% to 73%) was significantly different from the expected proportion of 50% (χ2i, 9.14, P= 0.002). Fifty-four women who had had previous pelvic surgery were excluded: total abdominal hysterectomy n= 25 (including 19 with uni-or bilateral salpingo-oophorectomy); unilateral salp-ingo-oophorectomy n= 4; ovarian cystectomy n= 1; myomectomy n= 1; caesarean section n= 1; laparotomy or laparoscopy with unspecified pelvic procedures n = 22, These exclusions did not, however, modify the over all results (left-sided lesions, 47/65,72%; 95% CI, 61% to 83%; χ2i, 12.9, P=0.001). Also including the bilateral ureteral endometriotic forms gave a total of 91/150 (61%) left-sided and 59/150 right-sided ureteral lesions.
The prevalence of ureteral endometriosis ranges from 0.01% to 1% of all women with the disease4. Consequently, it would be almost impossible to demonstrate significant asymmetry in lesion distribution based only on the evaluation of patients referred to a single centre. The findings in our large surgical series and the results of our systematic review of published cases confirm that, similar to ovarian endometriosis, ureteral disease is observed more frequently on the left than the right side. Interestingly, the proportion of left-sided gonadal and ureteral lesions is remarkably similar (63% and 64%).
A left ovarian endometrioma was present in each of our six patients, but this association was less consistent or not always verifiable in the cases reported in the literature. However, endometriosis of the ureter may not necessarily be secondary to endometriotic cysts only, but more generally to ectopic implantation of endometrial cells along the lateral gonadal aspect and ovarian fossa. Indeed, the asymmetry of both ovarian and ureteral forms may be the expression of a common underlying anatomical condition that facilitates adhesion and growth of endometrial cells on the left pelvic side wall. In our opinion, neither the coelomic metaplasia nor the embryonic cell rests theory can explain such a clear-cut difference in frequency distribution of ovarian and ureteral lesions between the two pelvic sides2,3.
Selection of women undergoing first-line abdomino-pelvic surgery among the cases reported in the literature should be more reliable for the identification of true instances of ureteral endometriosis, otherwise infiltrating disease or unsuspected injury at previous interventions might have been responsible for the lesion. The distribution of lesions remained unbalanced in favour of the left side also after exclusion of the patients operated before recognition of ureteral obstruction.
Distinction between the so-called intrinsic lesion type (endometriotic tissue within typically hyperplas-tic and fibrotic muscularis6) and the extrinsic lesion type (ureteral luminal narrowing caused by compression and/or fibrosis6) was not attempted because it must be based on histologic examination of a resected ureteral segment, which was not available in two of our patients. However, clinical presentation is similar and surgical exploration is always mandatory, so that differentiation may be more academic than practical. Furthermore, it cannot be excluded that the two forms constitute different degrees of the same histopatho-genetic process6.
A detailed description of the patients' symptoms and signs and of the diagnostic and treatment alternatives was not among the aims of the study. However, based on our findings, we would suggest urinary tract imaging in women with severe endometriotic parametrial infiltration or extensive rectovaginal lesions as even silent loss of renal function may occur. Pre-operative detection of obstructive uropathy allows adequate counselling about the type of surgery required and its potential sequelae, and timely scheduling of urologic consultation.
In conclusion, our study demonstrated a lateral asymmetry in the location of ureteral endometriosis, as almost two-thirds of the patients had ureteral lesions on the left-hand side. This is compatible with the menstrual reflux theory and with the anatomical differences of the left and right hemipelvis2,3.