Adenomyosis is believed to result from endometrial glands and stroma penetrating myometrium by lymphatic and vascular channels. Extent of disease is assessed by the distance of the deepest focus from the lower border of the endometrium. The minimum distance necessary for diagnosis has been debated, but ranges from 2 mm to more than 4 mm, or from one to two low-power fields. Another criterion for diagnosis has been the involvement of at least one third of myometrial thickness.1 None of those definitions has been accepted universally.
Etiology of adenomyosis is not known, but several pathophysiologic mechanisms have been suggested. The one supported by an animal experiment is the breakdown of the endometrial–myometrial border after trauma.1,2
Adenomyosis is identified in 20–30% of all uteri1; however, in specimens larger than 280 g it is much less common.3,4 This might be attributed to atrophic changes in the adenomyotic foci by large coexisting leiomyomas.3,4 Also, it is possible that small uteri with adenomyosis are extirpated for intractable symptoms, and therefore do not attain large size.
The published prevalence of adenomyosis ranges from 5–8% to 40–70% of all uterine specimens.1,5 That wide range stems from not only dissimilarities in uterine size in different studies, but also differences in ages of subjects and amounts of tissue sections examined. Another reason is a bias in most studies based on selected groups of women who had hysterectomies, and not on the general female population.1,2,5–8
Various histopathologic features were evaluated to assess severity of disease and correlate it to symptoms, including depth and number of foci, number of glands, number of glands per focus, and ratio of depth of foci to muscle thickness.5,6
Symptoms of adenomyosis are menorrhagia and dysmenorrhea, and they appear in 40–50% and 15–30% of women, respectively. One third of women are asymptomatic.1
The aims of this study were to assess prevalence of pregnancy terminations in women with and without adenomyosis, to evaluate the theory of endometrial trauma as a cause of it, and to assess the relation between depth and number of adenomyotic foci and their possible correlation to menorrhagia or dysmenorrhea. Uteri that weigh 280 g or more are different from smaller uteri in prevalence and microscopic appearance of adenomyosis, so we limited this study to specimens that weighed under 280 g.
Materials and Methods
This was a retrospective study of women who had hysterectomies for benign uterine enlargement between January 1994 and July 1997. Obstetric data, particularly histories of pregnancy terminations, and clinical information on menorrhagia and dysmenorrhea, were collected from patients' records.
Each pathologic specimen included the uterus and cervix. Five to ten paraffin-embedded hematoxylineosin–stained sections from each specimen, prepared and examined microscopically during surgery, were reevaluated by two observers. Adenomyosis and leiomyomas were identified by consensus, with no interobserver variability.
Adenomyosis was defined as endometrial glands or stroma within the myometrium at a depth of 2.5 mm or more. If adenomyosis was found, a detailed histopathologic evaluation included assessment of number of adenomyosis foci per slide and depth of adenomyosis foci within the myometrium, expressed as percentage of myometrial thickness, and graded as deep (exceeding 80%), intermediate (40–80%), and superficial (under 40%).
Women were stratified into four groups by histopathologic diagnosis: 17 had adenomyosis alone, 19 had adenomyosis with leiomyomas, 39 had leiomyomas alone, and 36 had neither.
Comparison of age, gravidity, parity, and incidence of spontaneous abortion, pregnancy termination, cesarean, and myomectomy was done by Fisher exact test or its generalization. Differences in continuous variables were examined by analysis of variance with paired differences assessed by Tukey's multiple comparison test. If the assumptions of analysis of variance were not met, the Kruskal–Wallis nonparametric test was substituted. Multinomial logistic regression analysis was used to analyze the likelihood of group membership according to pregnancy terminations, adjusting for differences in age, parity, and incidence of cesarean and D&C.
The correlation between number of adenomyosis foci and degree of myometrial depth, in the adenomyosis alone and adenomyosis with leiomyomas groups, was evaluated by regression analysis.
The Wilcoxon rank-sum test was used to compare the median number of adenomyosis foci in women with menorrhagia or dysmenorrhea to those without symptoms. Prevalence of menorrhagia and dysmenorrhea in women with various ranges of depths of foci was compared using chi square. In all statistical analyses, P < .05 was considered statistically significant.
The study included 111 women in whom uterine specimens weighed under 280 g. Thirty-six specimens had evidence of adenomyosis, a prevalence of 32.4%; none had evidence of lymphovascular invasion. Another group of 132 women with specimens that weighed 280 g or more, including eight with adenomyosis (6%), was not studied because the distortion of the uterine cavity by the leiomyomas and the lack of full-thickness sections of the uterine wall precluded accurate histopathologic evaluation.
Table 1 shows the mean age, gravidity, parity, and frequency of cesarean, D&C, myomectomy, spontaneous abortion, and pregnancy termination in the four groups. The gravidity in women with adenomyosis alone differed significantly from women with leiomyomas alone (P < .01).
Pregnancy termination occurred more than twice as frequently in women with adenomyosis alone or adenomyosis with leiomyomas than in women with leiomyomas alone and women with neither (P < .01).
Table 2 shows the multinomial logistic regression analysis adjusting for potential confounders. A pregnancy termination similarly increased risk for either adenomyosis alone or adenomyosis with leiomyomas. A pregnancy termination increased the odds ratio (OR) of a subsequent diagnosis of adenomyosis fourfold. The greater likelihood of adenomyosis among women with pregnancy terminations remained when women with leiomyomas alone were used as a reference group. Myomectomy was not included in Table 2 because only three women had it and the results were not changed when they were excluded.
The number of adenomyosis foci directly correlated with degree of myometrial depth of foci in specimens with adenomyosis alone (r = .46; P = .05) and adenomyosis with leiomyomas (r = .66; P < .001).
Dysmenorrhea was associated with the amount of adenomyotic foci. The median number in women with and without dysmenorrhea was 10 (range 2–30) and 4.5 (range 1–14), respectively, and differed significantly (P < .003). There was no significant difference between median numbers in women with menorrhagia (7, range 2–30) and without menorrhagia (7, range 1–17), respectively (P = .25).
Menorrhagia and dysmenorrhea were associated with degree of myometrial depth. Menorrhagia occurred in 36.8% of women with deep foci and 13.3% (P < .001) with intermediate foci, respectively. Dysmenorrhea was present in 77.8% of women with deep foci, compared with 12.5% with intermediate myometrial foci (P < .001). Neither menorrhagia nor dysmenorrhea was associated with superficially localized foci.
We used the currently accepted definition of one half of low-power field, which is equivalent to 2.5 mm3 to diagnose adenomyosis. It is diagnosed in 20–30% of uterine specimens1 and in 60–80% of women with it, the uterus is enlarged to up to 12 weeks' gestational size, equivalent to 280–320 g according to ACOG.10 The prevalence of 32.4% in our study agrees with the literature.1,5 However, our study was limited to specimens that weighed under 280 g because of infrequent diagnosis of adenomyosis in larger uteri, probably due to atrophy of foci by large leiomyomas3,4 and because sections collected from large specimens were not suitable for accurate evaluation.
The reason for penetration of glandular and stroma cells into the myometrium is not clear. In one theory, a breakdown of the myometrial–endometrial border, caused by trauma, is followed by reactive hyperplasia of the basalis and its penetration into the myometrium.1,2 In a recent case report,11 a laparoscopic myomectomy did not include reconstruction of layers by reapproximation of endometrium and myometrium. It resulted in adenomyosis 3 months later.
The concept of trauma as a cause of adenomyosis was supported by one animal study; removal of products of conception from one uterine horn by curettage and leaving an intact pregnancy in the contralateral horn resulted in adenomyosis of the emptied horn.2 The hypothesis was that vigorous curettage and high levels of estrogen, progesterone, and prolactin were important in the pathogenesis. The association we found between adenomyosis and history of pregnancy termination is compatible with that animal model and supports the notion that the procedure affects development of this disease. Other aspects in the reproductive histories of women in all four groups were similar except gravidity, which indicates the difference in number of pregnancy terminations. Our series consisted of very few individuals who had more than two terminations, so we were unable to draw conclusions about actual numbers of terminations.
According to statistics compiled by Azziz,1 35% of women with adenomyosis are asymptomatic, whereas others manifest with menorrhagia or dysmenorrhea.1,7,12 Improper uterine contractions during menses, increased endometrial surface, overproduction of prostaglandin,13 and hyperestrogenism1 are listed as causes of menorrhagia, whereas dysmenorrhea might be caused by uterine irritability or pseudodecidual edema around the adenomyosis foci.1
The relation between menorrhagia or dysmenorrhea and severity of disease, assessed by various histopathologic features, might be partially explained by adenomyosis with other abnormal uterine conditions. Coexisting conditions appear in 60–80% of adenomyosis cases and include leiomyomas, endometriosis, and salpingitis isthmica nodosa (35–55%, 5–20%, and 2–20%, respectively).5,7,12,14,15 Such a condition, rather than adenomyosis per se, might be the cause of symptoms. Dividing our subjects into pure and combined groups helped us show that adenomyosis might cause symptoms regardless of leiomyomas.
Our results showed no increase in number of foci in women with menorrhagia, but confirmed direct correlation of that number with dysmenorrhea, an observation made by Nishida6; therefore, we conclude that number of foci is definitely an important cause of dysmenorrhea.
Two studies showed direct correlation between prevalence of dysmenorrhea and depth of foci. Bird et al5 divided grades of penetration as grade 1, depth of one power field below the endometrium; grades 2 and 3, respectively, were foci localized within or beyond the limit of midmyometrium. Dysmenorrhea was reported by 4.3% of women with grade 1 compared with 42.4% with grade 2, and 83.3% with grade 3. In another study, Nishida6 found that depth of foci was significantly greater in 16 dysmenorrheic women than in 13 women without dysmenorrhea, and concluded that 80% penetration is a risk for dysmenorrhea, which we verified. It also shows a decrease in occurrence of dysmenorrhea with lesser degrees of myometrial depths and no dysmenorrhea in superficial disease. Unlike Nishida, we extended our histopathologic assessment and included menorrhagia in our analysis and found that menorrhagia was also related to depth of the foci within the myometrium, in contrast to a previous study.5
We found a direct correlation between the two histopathologic features and the clinical manifestations. We believe there is a causative relationship between those features that might imply that number and depth of foci are controlled by the same pathophysiologic mechanism. This study offers clinical evidence of the effect of previous pregnancy terminations on pathogenesis of this disease but does not provide data on the importance of actual numbers of abortions.