1 Service de Gynécologie Obstétrique II, Assistance PubliqueHôpitaux de Paris, Unité de Chirurgie Gynécologique, CHU Cochin, 123, bd Port-Royal, 75079 Paris Cedex 14, France and 2 Unité Inserm 149, Recherches Epidémiologiques en Santé Périnatale et Santé des Femmes, Paris, 123, bd Port-Royal, 75014 Paris, France
3 To whom correspondence should be addressed at: Unité Inserm 149, Recherches Epidémiologiques en Santé Périnatale et Santé des Femmes, 123, bd Port-Royal 75014 Paris, France. Email: afauconnier{at}chi-poissy-st-germain.fr
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: deep infiltrating endometriosis/diagnosis/dyspareunia/pelvic pain/questionnaires
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Treatment of posterior DIE may require extensive surgery (Redwine, 1992; Garry et al., 2000
), which may include USL resection (Chapron and Dubuisson, 1996
), partial colpectomy (Martin, 1988
; Donnez and Nisolle, 1995
) and resection of rectal endometriosis (Coronado et al., 1990
). Because these surgical procedures are time-consuming and require that patients receive specific information and appropriate preparation, presurgical diagnosis is very important.
Routine clinical examination cannot be used alone to diagnose DIE (Koninckx et al., 1996a; Chapron et al., 2002
). Magnetic resonance imaging is an interesting tool for presurgical diagnosis (Kinkel et al., 1999
), but its diagnostic accuracy has not been evaluated. Rectal endoscopic ultrasonography has excellent diagnostic value for posterior DIE (Ohba et al., 1996
; Fedele et al., 1998
). It requires bowel preparation as well as a skilled radiologist. Because its invasiveness, however, it may not be proposed to all women before laparoscopic evaluation and treatment of pelvic pain. A previous study of retrospective data (Fauconnier et al., 2002
) showed that the various specific anatomic locations of DIE are related to specific pain symptoms. These findings suggest that questioning may be helpful for predicting posterior DIE. That study, however, did not assess symptoms in a standardized fashion, with, for example, a question-naire. The aim of the present study was thus to evaluate prospectively the diagnostic accuracy of medical history and symptoms, collected with a standardized questionnaire, in predicting posterior DIE.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Each woman who agreed to participate was asked to complete a written self-administered questionnaire. A.F. and C.C. developed this questionnaire at our institution to evaluate endometriosis-related pain symptoms. The list of symptoms we included came from a comprehensive review of the literature and from our previous practice. Sentences were constructed from oral interviews during which we asked women who had pelvic pain symptoms and confirmed DIE to describe their pain. A draft questionnaire was tested on 12 women (three without and nine with pain) and modified to improve its comprehensibility. The questionnaire included items to evaluate the intensity and frequency of DM, dyspareunia and nonmenstrual pain, as well as the disability associated with each. The intensity of each painful symptom was evaluated according to either a 10 cm visual analogue scale (VAS) (Peveler et al., 1996) or a semiquantitative assessment (not at all; yes a little; yes a lot). Patients also recorded whether various symptoms (of the urinary or gastrointestinal tracts) occurred or increased during menses. The appendix includes several items from the questionnaire, which also asked for sociodemographic data, height, weight, body mass index, previous pelvic surgery, gravidity, parity, infertility (failure to conceive for more than 1 year), menstrual cycle characteristics and sexual activity. This questionnaire was given at the pre-operative consultation and completed during the month before before surgery.
Posterior DIE was diagnosed during laparoscopy (The American Fertility Society, 1993), which included thorough bidigital palpation of the retrocervical area under laparoscopic guidance. The diagnosis was based on the macroscopic appearance of the lesion, according to the following criteria: (1) obvious deep infiltration during excision (Koninckx and Martin, 1994
), requiring either USL resection, partial colpectomy or resection of the rectal wall to remove the implant; (2) visible nodule or infiltration associated with palpable induration (Martin et al., 1989
; Cook and Rock, 1991
; Reich et al., 1991
; Chapron and Dubuisson, 1996
; Possover et al., 2000
) at any of the following locations: torus uterinum, USL, posterior vaginal wall or anterior rectal wall at laparoscopy; (3) visible dark blue nodule of the posterior vaginal fornix on speculum examination (Vercellini et al., 1996
). In case of complete cul-de-sac obliteration, the surgeon had to perform a complete adhesiolysis to analyse the posterior area.
The diagnoses of the other patients were grouped as follows: (1) not endometriosis (e.g. ovarian cysts, tubal pathology, adhesions without evidence of endometriosis), (2) superficial endometriosis, (3) endometriomas (including one woman with anterior DIE).
Statistical methods
Women with posterior DIE were compared with the other patients. Pearson's 2 test was used for the qualitative variables and Student's t-test for the quantitative variables. The variables were dichotomized to simplify their use in constructing the diagnostic model as follows: the diagnostic accuracy of each possible cut-off point for the quantitative or ordinal variables was assessed by analyzing the receiver operating characteristics (ROC). Sensitivity, specificity, positive and negative likelihood ratios were calculated and the best cut-off points were considered to be those that maximized the sum of specificity and sensitivity.
Dichotomized variables associated with the presence of posterior DIE (at a threshold of P<0.10) were then entered into a multiple logistic regression model to select the best combination for predicting posterior DIE. A backward stepwise procedure was then used to remove variables so that the final model included only those variables independently associated with posterior DIE at the threshold of P<0.05. The model's predictive ability was then assessed by calculating the probability of posterior DIE for the overall study population, with its mathematical equation. Observed and predicted values were compared with the Hosmer and Lemeshow goodness-of-fit test (Hosmer and Lemeshow, 1989). We then calculated overall sensitivity, specificity and positive and negative likelihood ratios for the diagnostic value of the model.
The diagnostic model was then tested with the jackknife method (Efron and Gong, 1983); especially useful for assessing its stability. The method was applied as follows (Ezeh et al., 1999
): (1) the study population was randomly stratified into 10 subgroups of 13 (±1) women; (2) for each subgroup (i), a new prediction model (Mi) was constructed with the entire study sample except those in subgroup (i); (3) the model Mi was then used to predict the outcome of the women omitted in constructing it. These steps were repeated for each subgroup. All analyses were performed with Stata 6.0 software (Stata Corporation).
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Our study was specifically designed to evaluate the usefulness of questioning in diagnosing posterior DIE. Medical history and symptoms were assessed prospectively. Following the appropriate procedures (Falissard, 2001), we developed a questionnaire designed to assess pain symptoms related to endometriosis. The exhaustiveness of the symptoms studied was ensured by basing the list on a comprehensive literature review and on our previous practice (Fauconnier et al., 2002
). Finally, the study population was fairly large and included 51 women with posterior DIE.
This study nonetheless has several weaknesses. First, 30.1% of the women asked to participate did not answer the questionnaire. This may weaken the generalizability of the results. However, responders and non responders had similar age and similar diagnosis at laparoscopy.
The second limitation is that the diagnosis of posterior DIE was based on laparoscopic evaluation (The American Fertility Society, 1993) and not on histological results, as some authors recommend (Howard, 1993
; Redwine, 1999
). We chose to base inclusion of the women on macroscopic diagnosis of DIE because the posterior implants were not systematically removed during the study. Some women had DIE discovered incidentally at laparoscopy or did not want extensive surgery. Limiting inclusion to the cases which had laparoscopic removal of their posterior DIE would have led to indication bias. Nonetheless, the macroscopic criteria we used are very reliable for the diagnosis of posterior DIE (Martin et al., 1989
; Cornillie et al., 1990
; Koninckx and Martin, 1994
; Vercellini et al., 1996
; Possover et al., 2000
; Chapron et al., 2001
, 2002
; Fedele et al., 2001
; Bonte et al., 2002
).
Third, we performed no cross-validation studies on independent samples. There is accordingly a risk that we overfit the data when we developed the prediction model (Altman and Royston, 2000). To safeguard against overfitting, we used the jackknife procedure to test for instability of the model. This method is known to yield results similar to those obtained by testing the model in an independent population (Rothpearl, 1989
; Ezeh et al., 1999
). Nonetheless, validation studies of samples from different centres are important for evaluating the actual predictive ability of our diagnostic model (Coste et al., 1997
). Because the prevalence of posterior DIE in our population was particularly high, it is important to verify the diagnostic accuracy of our model in a population with a lower prevalence of DIE. Its negative predictive value will certainly increase.
Painful defecation during menses was the strongest predictor of posterior DIE. This result is consistent with previous series that report significant reductions in painful defecation after surgical (Anaf et al., 2001) or medical treatment (Fedele et al., 2001
) of posterior DIE. In a previous retrospective study we found that this symptom was specifically related to infiltration of the posterior vaginal wall (Fauconnier et al., 2002
) rather than other DIE locations. In the present study, painful defecation during menses was present in more than half the women with posterior DIE, regardless of the anatomic location of the posterior implant. This finding suggests that posterior DIE is to be viewed more as a continuum of possible locations (e.g. USL, vagina, rectum) than a specific and distinct location.
Although dyspareunia is a well known marker for posterior DIE (Vercellini et al., 1996; Chapron et al., 1999
), some form of dyspareunia is also reported frequently in general populations (Jamieson and Steege, 1996
) and in women with any type of endometriosis (Mahmood et al., 1991
; Fedele et al., 1992
; Al-Badawi et al., 1999
). With the use of a standardized questionnaire, we found that only severe dyspareunia (i.e. VAS
8 cm, apareunia, or always having dyspareunia) and not dyspareunia itself (regardless of its severity) was linked to posterior DIE. Another prospective study (Porpora et al., 1999
) compared DIE involving USL with other subtypes of endometriosis and found dyspareunia to have a discriminant value, at the same cut-off point we used.
Surprisingly, no items evaluating DM correlated with the presence of posterior DIE. Most of our controls (some of whom had other subtypes of endometriosis) had similar or higher levels of DM. The severity of DM was in fact strongly related to the presence of any subtype of endometriosis (data not shown), but not to any particular subtype. These findings are consistent with those by Vercellini et al. (1996).
Apart from ours, two studies have previously developed diagnostic models for the presurgical prediction of endometriosis, based on noninvasive tools. One (Koninckx et al., 1996a) developed a diagnostic model to predict deep endometriosis based on two criteria: menstrual examination and CA-125 assay during the follicular phase. This model, with a sensitivity of 83% and a specificity of 87%, had a far better diagnostic value than ours. It has however two important flaws: the study includes only 13 women with DIE lesions; and, contrary to our model, it did not distinguish women with DIE from those with endometriomas. The other study (Eskenazi et al., 2001
) was specifically designed to assess the prediction of endometriosis with the use of medical history, clinical examination and transvaginal ultrasonography. Apart from endometriomas, which were accurately predicted by transvaginal ultrasonography, no other subtypes of endometriosis were reliably predicted by any of the variables tested. The originality of our diagnostic model is that it was specifically designed to predict posterior DIE among women with chronic pelvic pain symptoms rather than to predict endometriosis of any subtype, as the other diagnostic models do.
Physical examination, including palpation of the retrocervical area with manual examination and searching for a dark blue nodule at speculum examination is important in assessing women with suspected DIE. However, routine physical examination was not included in the study because it has insufficient diagnostic value for screening women with posterior DIE. Indeed, >50% of women presenting with posterior DIE may have negative or non specific findings at physical examination (Koninckx et al., 1996a; Chapron et al., 2002
). On the other hand, physical examination performed during menstruation was found to have interesting diagnostic value (sensibility: 77% and specificity 76%) to diagnose either deep endometriosis, cystic ovarian endometriosis, or cul-de-sac obliteration (Koninckx et al., 1996a
). However, physical examination in the context of endometriosis requires skilled practitioners and inter-observer variability may be quite high, contrarily to our diagnostic model whose reproducibility is guaranteed by the fact that it is based on a standardized questionnaire.
Compared with more invasive diagnostic tools, such as diagnostic laparoscopy (Martin et al., 1989; Cook and Rock, 1991
), rectal endoscopic ultrasonography (Ohba et al., 1996
; Fedele et al., 1998
) or magnetic resonance imaging (Bazot et al., 2004
), the diagnostic accuracy of our model can best be categorized as rather moderate. However, it may be appropriately combined with other noninvasive diagnostic tools, such as menstrual examination (Koninckx et al., 1996a
), CA 125 sampling (Koninckx et al., 1996a
), or transvaginal ultrasonography (Bazot et al., 2003
) to improve the pre-operative detection of posterior DIE. Further studies are needed to evaluate the accuracy of these strategies.
We recommend our diagnostic model's use to help clinicians dealing with endometriosis to screen women who may require more invasive exploration before laparoscopic evaluation and treatment of pelvic pain. The aim of this policy is to promote a one-step surgical procedure for the complete removal of the DIE. Rectosigmoid involvement is rather frequent in women with posterior DIE and may lead the surgeon to perform extensive surgical procedures, including bowel resection (Chapron et al., 2003). Adequate pre-operative evaluation, including rectal endoscopic ultrasonography (Schroder et al., 1997
; Bazot et al., 2003
; Chapron et al., 2004
), magnetic resonance imaging (Kinkel et al., 1999
; Bazot et al., 2004
; Chapron et al., 2004
) profile rectal radiography with barium enema (Donnez et al., 1997
), or rectosigmoidoscopy (Anaf et al., 2001
) is thus necessary to evaluate the extension of disease. The need for detection of women with posterior DIE before surgery is also important to provide adequate counselling (Garry et al., 2000
). Excision of posterior DIE is technically demanding and may be associated with significant surgical risks (Koninckx et al., 1996b
). Women must thus be advised of the risks of laparotomy and colostomy (Garry et al., 2000
).
![]() |
Acknowledgements |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Altman DG and Royston P (2000) What do we mean by validating a prognostic model? Stat Med 19, 453473.[CrossRef][ISI][Medline]
Anaf V, Simon P, el Nakadi I et al. (2001) Impact of surgical resection of rectovaginal pouch of douglas endometriotic nodules on pelvic pain and some elements of patients' sex life. J Am Assoc Gynecol Laparosc 8, 5560.[ISI][Medline]
Bazot M, Darai E, Hourani R et al. (2004) Deep pelvic endometriosis: MR imaging for diagnosis, prediction of extension of disease. Radiology 232, 379389.
Bazot M, Detchev R, Cortez A et al. (2003) Transvaginal sonography and rectal endoscopic sonography for the assessment of pelvic endometriosis: a preliminary comparison. Hum Reprod 18, 16861692.
Bonte H, Chapron C, Vieira M et al. (2002) Histologic appearance of endometriosis infiltrating uterosacral ligaments in women with painful symptoms. J Am Assoc Gynecol Laparosc 9, 519524.[ISI][Medline]
Chapron C and Dubuisson JB (1996) Laparoscopic treatment of deep endometriosis located on the uterosacral ligaments. Hum Reprod 11, 868873.[Abstract]
Chapron C, Dubuisson JB, Fritel X et al. (1999) Operative management of deep endometriosis infiltrating the uterosacral ligaments. J Am Assoc Gynecol Laparosc 6, 3137.[ISI][Medline]
Chapron C, Dubuisson JB, Pansini V et al. (2002) Routine clinical examination is not sufficient for diagnosing and locating deeply infiltrating endometriosis. J Am Assoc Gynecol Laparosc 9, 115119.[ISI][Medline]
Chapron C, Fauconnier A, Vieira M et al. (2003) Anatomical distribution of deeply infiltrating endometriosis: surgical implications and proposition for a classification. Hum Reprod 18, 157161.
Chapron C, Jacob S, Dubuisson JB et al. (2001) Laparoscopically assisted vaginal management of deep endometriosis infiltrating the rectovaginal septum. Acta Obstet Gynecol Scand 80, 349354.[CrossRef][ISI][Medline]
Chapron C, Vieira M, Chopin N et al. (2004) Accuracy of rectal endoscopic ultrasonography and magnetic resonance imaging in the diagnosis of rectal involvement for patients presenting with deeply infiltrating endometriosis. Ultrasound Obstet Gynecol 24, 175179.[CrossRef][ISI][Medline]
Cook AS and Rock JA (1991) The role of laparoscopy in the treatment of endometriosis. Fertil Steril 55, 663680.[ISI][Medline]
Cornillie FJ, Oosterlynck D, Lauweryns JM et al. (1990) Deeply infiltrating pelvic endometriosis: histology and clinical significance. Fertil Steril 53, 978983.[ISI][Medline]
Coronado C, Franklin RR, Lotze EC et al. (1990) Surgical treatment of symptomatic colorectal endometriosis. Fertil Steril 53, 411416.[ISI][Medline]
Coste J, Bouyer J and Job-Spira N (1997) Construction of composite scales for risk assessment in epidemiology: an application to ectopic pregnancy. Am J Epidemiol 145, 278289.[Abstract]
Donnez J and Nisolle M (1995) Advanced laparoscopic surgery for the removal of rectovaginal septum endometriotic or adenomyotic nodules. Baillieres Clin Obstet Gynaecol 9, 769774.[ISI][Medline]
Donnez J, Nisolle M, Gillerot S et al. (1997) Rectovaginal septum adenomyotic nodules: a series of 500 cases. Br J Obstet Gynaecol 104, 10141018.[ISI][Medline]
Efron B and Gong G (1983) A leisurely look at the bootstrap, the jackknife and cross validation. Am Statist 37, 3648.[ISI]
Eskenazi B, Warner M, Bonsignore L et al. (2001) Validation study of nonsurgical diagnosis of endometriosis. Fertil Steril 76, 929935.[CrossRef][ISI][Medline]
Ezeh UI, Taub NA, Moore HD et al. (1999) Establishment of predictive variables associated with testicular sperm retrieval in men with non-obstructive azoospermia. Hum Reprod 14, 10051012.
Falissard B (ed) (2001) Mesurer la subjectivité en santé. In Perspectives méthodologiques et statistiques. Masson, Paris.
Fauconnier A, Chapron C, Dubuisson JB et al. (2002) Relation between pain symptoms and the anatomic location of deep infiltrating endometriosis. Fertil Steril 78, 719726.[CrossRef][ISI][Medline]
Fedele L, Bianchi S, Bocciolone L et al. (1992) Pain symptoms associated with endometriosis. Obstet Gynecol 79, 767769.[Abstract]
Fedele L, Bianchi S, Portuese A et al. (1998) Transrectal ultrasonography in the assessment of rectovaginal endometriosis. Obstet Gynecol 91, 444448.
Fedele L, Bianchi S, Zanconato G et al. (2001) Use of a levonorgestrel-releasing intrauterine device in the treatment of rectovaginal endometriosis. Fertil Steril 75, 485488.[CrossRef][ISI][Medline]
Garry R, Clayton R and Hawe J (2000) The effect of endometriosis and its radical laparoscopic excision on quality of life indicators. BJOG 107, 4454.[Medline]
Hosmer DW and Lemeshow S (eds) (1989) Applied logistic regression. Wiley & Sons, New York.
Howard FM (1993) The role of laparoscopy in chronic pelvic pain: promise and pitfalls. Obstet Gynecol Surv 48, 357387.[Medline]
Jamieson D and Steege J (1996) The prevalence of dysmenorrhea, dyspareunia, pelvic pain, and irritable bowel syndrome in primary care practices. Obstet Gynecol 87, 5558.
Kinkel K, Chapron C, Balleyguier C et al. (1999) Magnetic resonance imaging characteristics of deep endometriosis. Hum Reprod 14, 10801086.
Koninckx P, Meuleman C, Oosterlynck D et al. (1996a) Diagnosis of deep endometriosis by clinical examination during menstruation and plasma CA-125 concentration. Fertil Steril 65, 280287.[ISI][Medline]
Koninckx PR and Martin D (1994) Treatment of deeply infiltrating endometriosis. Curr Opin Obstet Gynecol 6, 231241.[ISI][Medline]
Koninckx PR, Meuleman C, Demeyere S et al. (1991) Suggestive evidence that pelvic endometriosis is a progressive disease, whereas deeply infiltrating endometriosis is associated with pelvic pain. Fertil Steril 55, 759765.[ISI][Medline]
Koninckx PR, Timmermans B, Meuleman C et al. (1996b) Complications of CO2-laser endoscopic excision of deep endometriosis. Hum Reprod 11, 22632268.[Abstract]
Mahmood TA, Templeton AA, Thomson L et al. (1991) Menstrual symptoms in women with pelvic endometriosis. Br J Obstet Gynaecol 98, 558563.[ISI][Medline]
Martin DC (1988) Laparoscopic and vaginal colpotomy for the excision of infiltrating cul-de-sac endometriosis. J Reprod Med 33, 806808.[ISI][Medline]
Martin DC and Batt RE (2001) Retrocervical, retrovaginal pouch, and rectovaginal septum endometriosis. J Am Assoc Gynecol Laparosc 8, 1217.[ISI][Medline]
Martin DC, Hubert GD and Levy B (1989) Depth of infiltration of endometriosis. J Gynecol Surg, 55.
Ohba T, Mizutani H, Maeda T et al. (1996) Evaluation of endometriosis in uterosacral ligaments by transrectal ultrasonography. Hum Reprod 11, 20142017.[Abstract]
Peveler R, Edwards J and Daddow J (1996) Psychosocial factors and chronic pelvic pain: a comparison of women with endometriosis and with unexplained pain. J Psychosomat Res 40, 305315.[CrossRef][ISI][Medline]
Porpora MG, Koninckx PR, Piazze J et al. (1999) Correlation between endometriosis and pelvic pain. J Am Assoc Gynecol Laparosc 6, 429434.[ISI][Medline]
Possover M, Diebolder H, Plaul K et al. (2000) Laparascopically assisted vaginal resection of rectovaginal endometriosis. Obstet Gynecol 96, 304307.
Redwine DB (1992) Laparoscopic en bloc resection for treatment of the obliterated cul-de-sac in endometriosis. J Reprod Med 37, 695698.[ISI][Medline]
Redwine DB (1999) Ovarian endometriosis: a marker for more extensive pelvic and intestinal disease. Fertil Steril 72, 310315.[CrossRef][ISI][Medline]
Reich H, McGlynn F and Salvat J (1991) Laparoscopic treatment of cul-de-sac obliteration secondary to retrocervical deep fibrotic endometriosis. J Reprod Med 36, 516522.[ISI][Medline]
Rothpearl A (1989) The jackknife technique in statistical analysis. Chest 95, 940.
Schroder J, Lohnert M, Doniec JM et al. (1997) Endoluminal ultrasound diagnosis and operative management of rectal endometriosis. Dis Colon Rectum 40, 614617.[ISI][Medline]
The American Fertility Society (1993) Management of endometriosis in the presence of pelvic pain. Fertil Steril 60, 952955.[ISI][Medline]
Vercellini P, Trespidi L, de Giorgi O et al. (1996) Endometriosis and pelvic pain: relation to disease stage and localization. Fertil Steril 65, 299304.[ISI][Medline]
Submitted on May 26, 2004; accepted on October 29, 2004.