Best Practice & Research Clinical Obstetrics and Gynaecology Vol. 20, No. 4, pp. 465e477, 2006 doi:10.1016/j.bpobgyn.2006.01.017 available online at http://www.sciencedirect.com
2 Adenomyosis: epidemiological factors Paolo Vercellini*
MD
Associate Professor
Paola Vigano`
PhD
Research Biologist
Edgar Somigliana
MD, PhD
Clinical Assistant
Raffaella Daguati
MD
Research Fellow
Annalisa Abbiati
MD
Research Fellow Department of Obstetrics and Gynaecology, University of Milan, and the Centre for Research in Obstetrics and Gynaecology, Milan, Italy
Luigi Fedele
MD
Professor Department of Obstetrics and Gynaecology, University of Milan, Milan, Italy
Epidemiological studies of adenomyosis are difficult to interpret because the diagnostic criteria vary so widely that the disease may be easily over-diagnosed. This would severely hamper any attempt to define incidence and prevalence of the condition and the related risk factors, and would limit the possibility of clarifying to what extent adenomyosis contributes to clinical symptoms. There is a need for stringent and widely accepted diagnostic criteria in order to define not only the presence of adenomyosis but also depth of penetration and degree of spread of foci. Moreover, the evidence available on epidemiological characteristics of women with adenomyosis is greatly biased by the type of population studied, i.e. women undergoing hysterectomy. Therefore, a consensus on non-surgical diagnostic criteria at transvaginal ultrasonography and MRI is
* Corresponding author. Tel./Fax: þ39 0257 992331. E-mail address:
[email protected] (P. Vercellini). 1521-6934/$ - see front matter ª 2006 Elsevier Ltd. All rights reserved.
466 P. Vercellini et al
indispensable and urgently needed in order to be able to conduct epidemiological studies in women younger than those evaluated until now. Key words: adenomyosis; menhorragia; pelvic pain; infertility; hysterectomy.
Adenomyosis is a common gynaecological disorder characterized by the presence of endometrial glands and stroma deep within the myometrium associated with myometrial hypertrophy and hyperplasia.1 Although it has been recognized for over a century, reliable epidemiological studies on this condition are limited, probably because in the past only postoperative diagnoses were possible.2e4 The recent development of new diagnostic techniques (i.e. transvaginal ultrasonography and MRI) seems to have overcome this impasse, generating interest in a disease which has been on the one hand misunderstood and on the other abused because of its facile preoperative diagnosis.5e8 Moreover, the aetiology of adenomyosis is still unclear, as are its prevalence and incidence in the general population, its natural history, the factors that cause progression of the lesions, and its relationship with infertility.2e4 In the present overview we analyse the most recent epidemiological findings, with the aim of defining the frequency of adenomyosis and the risk factors associated with its development. Only articles published in the English language literature in the last 15 years have been considered, because prior to this time period diagnostic criteria were extremely variable and adequate imaging techniques were not always available. DEFINITION OF ADENOMYOSIS It is generally agreed that adenomyosis occurs when the normal boundary between the endometrial basal layer and the myometrium is disrupted.9 As a consequence of this disruption, the endometrial glands invade the myometrium, resulting in ectopic intramyometrial glands which cause adjacent myometrial hypertrophy and hyperplasia. What triggers this process is unclear.2e4 Islands of adenomyosis may be scattered throughout the uterine musculature, giving origin to the diffuse form of the disease, or, less frequently, they occur in a localized, focal form, the so-called adenomyoma.1,10 The ectopic mucosa resembles non-secretory basal endometrium, and a direct connection between the basalis portion of the endometrium and heterotopic foci has been demonstrated.11 The posterior myometrial wall is usually involved to a greater extent than the other uterine portions.1,10 By tradition, a histological diagnosis was made when endometrial glands and stroma were found at least one low-power field beneath the endomyometrial junction (4 mm),3 but even a two low-power-field distance was proposed.12 However, less stringent criteria have been adopted recently. Zaloudek and Norris10 indicate that adenomyosis should be diagnosed when the distance between the lower border of the endometrium and the affected myometrial area is over one half of a low-power field (w2.5 mm). Adenomyosis ‘sub-basalis’ has been defined as minimal glandular invasion extending 25% of the
Adenomyosis: epidemiological factors 467
total thickness of the myometrium. During the past 15 years the vast majority of researchers have used just the 2.5 mm cut-off limit and, as a consequence, this distance between the endomyometrial junction and the affected area has become the standard agreed depth of invasion for the diagnosis. Several schemes have been suggested to grade adenomyosis based on lesion type and extension. This seems important because symptomatology and response to conservative treatments have been shown to correlate with degree of myometrial penetration and extent of involvement.16 Siegler and Camilien14 suggested classifying adenomyosis according to the depth of myometrial penetration. Grades 1, 2, and 3 correspond, respectively, to adenomyotic involvement of the inner third (superficial adenomyosis), two thirds, and entire myometrium (deep adenomyosis). Furthermore, adenomyosis should also be graded as mild, moderate, or severe according to the number of endometrial islets observed at histological examination (1e3, 4e9, and 10, respectively). Hulka et al17 considered adenomyosis as mild when only microscopic foci were present or only the inner third of the myometrium was involved with disease. Focal disease included specimens with focal adenomyomas, and severe or diffuse adenomyosis was diagnosed when the specimen showed disease extending into the outer two thirds of the myometrium and up to gross involvement of the entire uterus. Sammour et al16 classified adenomyosis observed in hysterectomy specimens into four categories, based on the degree of penetration: 75% of myometrial thickness. Furthermore, they calculated the ‘penetration ratio’ (depth of penetration/myometrial thickness) representing the extent of the disease. The degree of spread was assessed by studying the number of foci per slide. We recommend following stringent diagnostic criteria to assess the depth of myometrial penetration by glands and stroma both in terms of absolute measurement (with exclusion of sub-basal lesions) and of proportion of uterine wall thickness invasion. Moreover, a grading system should be adopted in order to describe the degree of intramyometrial lesion extension. This could avoid over-diagnosing the condition or evaluating potentially very different study populations, and might enable reliable comparisons of more consistent data by researchers and clinicians. Based on the proposal of Siegler and Camilien,14 four parameters should be indicated to describe adenomyosis: i.e. presence (>2.5 mm from the endomyometrial junction), depth of penetration (up to one third, mild disease; between one and two thirds, moderate disease; more than two thirds, severe disease), degree of spread of the condition in terms of number of foci per low-power field (1e3 islets, grade 1 disease; 4e10 islets, grade 2 disease; >10 islets, grade 3 disease), and configuration of lesion (diffuse versus nodular/focal). PREVALENCE OF ADENOMYOSIS The reported prevalence of adenomyosis in available surgical series varies widely based on the histological criteria adopted for diagnosis and the technique used to obtain myometrial samples (Table 1). In order to evaluate the prevalence and risk factors for adenomyosis, data of 1334 consecutive women undergoing hysterectomy between 1990 and 1992 at the Department of Obstetrics and Gynaecology of the University of Milan were analysed.18 Adenomyosis was diagnosed in 332 cases (25%). The condition was present in 146 of the 627 patients (23%) with fibroids and menorrhagia, 68 of the 265 (26%) with prolapse, 21 of the 98 (21%) with ovarian cysts, 19 of the 100 (19%) with cervical cancer,
468 P. Vercellini et al
Table 1. Estimated prevalence of histologically confirmed adenomyosis in surgical series published in the period 1990e2004.a Authors 39
Shaikh et al Chrysostomou et al35 McCausland et al40,c Vercellini et al18 Seidman et al21 Vavilis et al19 Vercellini et al41,d Parazzini et al20 Vercellini et al42 Whitted et al24 Levgur et al23 Bergholt et al22 Sammour et al16 Curtis et al25 Panganamamula et al26 a b c d
Year
Number of patients
Prevalence (%)
95% CI (%)b
1990 1991 1992 1995 1996 1996 1996 1997 1998 2000 2000 2001 2002 2002 2004
419 646 50 1334 1252 394 72 707 102 200 111 549 94 1850 873
57 25 66 25 39 20 18 21 28 32 32 14 16 20 47
52e61 22e29 51e79 23e27 36e42 16e23 10e30 18e24 20e38 26e39 24e42 11e17 9e25 18e22 44e51
Hysterectomy was the surgical procedure performed unless otherwise specified. 95% confidence intervals (CI) based on Poisson’s approximation. Myometrial resectoscopic biopsy at operative hysteroscopy. Myometrial needle-biopsy at laparoscopy.
31 of the 110 (28%) with endometrial cancer, 16 of the 57 (28%) with ovarian cancer, and 19 of the 77 (25%) with miscellaneous indications. The above differences were not statistically significant. A very similar study was conducted by Vavilis et al19 on 594 women undergoing hysterectomy at the Department of Obstetrics and Gynaecology of the University of Thessaloniki. Adenomyosis was found in 116 patients (20%). The condition was present in 63 patients with fibroids (21%), 11 with genital prolapse (26%), 11 with benign ovarian tumours (18%), six with endometrial hyperplasia (14%), two with cervical cancer (18%), 10 with endometrial cancer (16%), and 13 with ovarian cancer (21%). In a second epidemiological study conducted at the University of Milan,20 adenomyosis was identified in 150/707 women (21%). The frequency of the condition was similar in women with indication for surgery for ovarian cysts (30%) and prolapse (31%), but was lower in women with fibroids and menorrhagia (15%) or genital cancer (25%). Seidman and Kjerulff reviewed 1252 pathology reports on hysterectomy specimens from women enrolled in the Maryland Women’s Health Study in order to assess the variability in the frequency of histological diagnosis of adenomyosis among 33 hospitals within the same geographic region.21 The frequency of adenomyosis ranged from 12% to 58% among the different hospitals, and from 10% to 88% among the 25 different pathologists. These major variations could not be explained by differences in patient age or number of pregnancies. Based on these findings, the authors suggest that adenomyosis may be over-diagnosed, and that stringent and standard diagnostic criteria in practice are needed. The wide variability in the reported frequency of adenomyosis has been confirmed by Bergholt et al in a study including 549 consecutive women undergoing hysterectomy over a 2-year period.22 The prevalence of adenomyosis varied from 10 to 18% depending on different diagnostic criteria.
Adenomyosis: epidemiological factors 469
Table 2. Risk factors associated with adenomyosis. Risk factor Menstrual and reproductive factors:
Direction and consistency of effect Age at menarche Number of births Spontaneous abortion Menopausal status at intervention Presence of endometriosis OC and IUD use Menorrhagia Dysmenorrhoea Chronic pelvic pain Dyspareunia Infertility
e limited data [[ consistent [[ consistent e limited data
Health habits:
Smoke
Y limited data
Surgical trauma:
Age at surgery Indication for surgery D&C Induced abortion Caesarean section
e consistent e consistent [ consistent [ inconsistent [ inconsistent
Endometrial hyperplasia and endometrial cancer:
Endometrial hyperplasia Endometrial cancer
[ consistent e limited data
[ inconsistent e limited data [ consistent [ inconsistent [ inconsistent e limited data [ inconsistent
OC, oral contraceptive; IUD, intrauterine device; D&C, dilation and curettage.
DETERMINANTS OF ADENOMYOSIS Several risk factors for adenomyosis have been evaluated. However, few studies have been correctly designed and even fewer adequately powered to identify moderate increases in relative risk (Table 2). Menstrual and reproductive factors In the first Milan epidemiological study on adenomyosis18 the frequency of the condition was directly associated with the number of births and tended to be higher in subjects with spontaneous and induced abortions. In fact, the odds ratio of the disease in women with one and more than two births was 1.3 and 1.5, respectively, in comparison with nulliparous women (c2 trend 5.76, P < 0.05). No relationship emerged between risk of adenomyosis and age at surgery, age at menarche, indication for surgery, menopausal status at intervention, and presence of endometriosis. According to these findings, adenomyosis does not seem to be related to particular clinical conditions except parity. In the Thessaloniki study19 a significant association was observed not only between adenomyosis and parity (OR ¼ 5.03) but also with caesarean section, induced abortion, dysmenorrhoea, and abnormal uterine bleeding, whereas no relationship was found between adenomyosis and endometriosis.
470 P. Vercellini et al
With the aim of analysing risk factors for adenomyosis, a second, cross-sectional study was conducted on 707 consecutive women who underwent hysterectomy at the Department of Obstetrics and Gynaecology of the University of Milan between 1993 and 1994.20 The mean age at hysterectomy was 53 11 years in women with adenomyosis and 48 14 in those without the condition (P < 0.05). The frequency of adenomyosis was higher in parous women: in comparison with nulliparae the odds ratios of the disease were 1.8 (95% CI 0.9e3.4) and 3.1 (95% CI 1.7e5.5) respectively in women reporting one and more than two births (c2 trend 20.71, P < 0.01). No relationship emerged with age at first birth, history of induced abortion, the use of oral contraceptives or an intrauterine device, and presence of endometriosis. The risk of adenomyosis was higher in women self-reporting heavy flows in comparison with those reporting regular flows (odds ratio 1.7, 95% CI 1.1e2.6). Age at menarche, menopausal status, lifelong menstrual pattern, and duration of flows were not associated with the frequency of adenomyosis. Women with the condition tended to report non-menstrual pelvic pain and dyspareunia more frequently, but the findings were not significant. Dysmenorrhoea was not related to the frequency of adenomyosis. Bergholt et al22 did not observe any significant association between adenomyosis and dysmenorrhoea, dyspareunia, chronic pelvic pain, indication for hysterectomy including menorrhagia, age at surgery, and parity. Health habits In the second Milan epidemiological study20 women who smoked tended to be at decreased risk of adenomyosis; in comparison with women who had never smoked, the risk for current smokers was 0.7 (0.3e1.3) and the risk decreased with number of cigarettes smoked per day, the odds ratios being 0.8 and 0.6 respectively in women reporting 10 cigarettes smoked per day (c2 trend 3.57, P ¼ 0.06). Surgical trauma In the second Milan series,20 women who underwent dilatation and curettage had an odds ratio of adenomyosis of 2.2 (95% CI 1.4e4.0). In order to verify the theory of endometrial trauma as a cause of adenomyosis, Levgur et al23 evaluated the frequency of prior pregnancy terminations in 111 women with and without adenomyosis who underwent hysterectomy. A total of 10/17 (59%) women with adenomyosis and 9/19 (47%) with adenomyosis plus leiomyomas underwent previous pregnancy terminations compared with 8/39 (20%) of those with leiomyomas alone and 8/36 (22%) of those with neither condition (P < 0.01), suggesting a role of surgical trauma in the pathogenesis of endometrial invasion of the myometrium. In a retrospective study including 200 women who underwent hysterectomy for benign disease, Whitted et al observed an increased prevalence of adenomyosis in subjects with prior caesarean delivery.24 Of the 64 women (32%) found to have adenomyosis, 19 (30%) had a history of caesarean delivery, whereas the proportion of women who delivered by caesarean section in the group without adenomyosis was 23% (25/111; OR 1.87, 95% CI 0.94e3.74). Curtis et al analysed surgical disruption of the endomyometrial border by sharp curettage as a risk factor for adenomyosis in a cohort of 1850 women undergoing
Adenomyosis: epidemiological factors 471
hysterectomy during a 4-year period.25 The 368 women (20%) with adenomyosis reported a history of induced abortion more frequently with respect to women without adenomyosis (17% and 13%, respectively; P ¼ 0.02). The prevalence of caesarean delivery and curettage in the non-pregnant status (D&C) was similar between the two groups. Women who were more likely to have had pregnancies terminated by sharp curettage had an increasing trend in risk with increasing number of abortions, with an odds ratio of 15.5 (95% CI 1.7e138.2) for more than three abortions. For women who were more likely to have had pregnancies terminated by suction evacuation, those with more than three abortions had an increased risk of adenomyosis of 5.9 (95% CI 1.5e23.3). The authors suggest that repeated sharp curettage during pregnancy may greatly increase the risk of adenomyosis by disrupting the endometriale myometrial border and facilitating implantation, embedding, and survival of endometrium within the myometrial wall. Interestingly, sharp curettage in the non-pregnant status did not increase the risk. Panganamamula et al observed adenomyosis in 412 of 873 women (47%) who underwent hysterectomy for benign conditions during an 8-year period.26 When individual types of surgery were analysed separately, no significant differences emerged between women with and without adenomyosis with regard to prior caesarean delivery, myomectomy, endometrial ablation, dilatation and evacuation, and D&C. However, after pooling of all procedures, history of any prior uterine surgery significantly increased the risk of adenomyosis (49% versus 41%; OR 1.37, 95% CI 1.05e1.79). At odds with the above findings, in the series of Bergholt et al no significant association emerged between adenomyosis and previous caesarean section, endometrial curettage or uterine evacuation.22
Depth of penetration and degree of spread In order to assess the possible relationship of depth and number of adenomyotic foci with menorrhagia and dysmenorrhoea, Levgur et al expressed myometrial invasion as a percentage of uterine wall thickness, graded as deep (>80%), intermediate (40e 80%), and superficial (
