Reprinted with permission
The objective of this manuscript is to study the pregnancy rate and the prognostic factors for conception after myomectomy carried out in cases of infertility.
In our opinion, post myomectomy adhesions could have an adverse effect on fertility after myomectomy.
Myomas responsible for menometrorrhagia are also responsible for infertility. In the presence of an associated male, tubal or ovulatory factor, the results were poor and it is difficult to say if a myomectomy should be performed in these cases in order to enhance fertility.
It is common for a uterine myoma to be discovered during the work-up for infertility, for which myomectomy has been recommended for many years past (Bonney, 1931; Miller and Tirone, 1933). Although the efficiency of myomectomy in restoring fertility has never been proved by a random clinical trial, there are several arguments indicating that myomas bear some responsibility for infertility: (1) there is a lower pregnancy rate in women presenting a myoma (Parazzini et al., 1988; Ross et al., 1986); (2) the association between myoma and infertility has been reported by two observational surveys (Marshall et al., 1998; Parazzini et al., 1996); (3) the rate of implantation for pregnancy obtained by IVF is lower in women with intramural myomas (Eldar-Geva et al., 1998; Stovall et al., 1998).
Myomectomy carried out in a context of infertility can be the origin of adverse effects on the reproductive function: post operative adnexal adhesions (Dubuisson et al., 1998; Tulandi et al., 1993; Ugur et al., 1996); uterine rupture after myomectomy carried out by laparotomy (Georgakopoulos and Bersis, 1981; Golan et al., 1990; Ozeren et al., 1997; Palerme and Friedman, 1966), by laparoscopy (Dubuisson et al., 1995; Friedmann et al., 1996; Harris, 1992), or by hysteroscopy (Abbas and Irvine, 1997). Systematic use of myomectomy for infertility is a subject of debate, in particular in case of intramural or sub serous myoma, those of medium size or without symptoms (Berkeley et al., 1983; Paulson, 1993; Vollenhoven et al., 1990). Some authors recommend it should be carried out when no other infertility factor has been found (Buttram and Reiter, 1981; Dubuisson et al., 2000; Vercellini et al., 1998; Verkauf, 1992) but do not state which characteristics of the myomas (size, number, depth of penetration into the myometrium) justify operation. When there are infertility factors associated with the presence of myomas the benefit observed after myomectomy remains unclear (Dubuisson et al., 2000; Vercellini et al., 1998). Knowledge of the prognostic factors which influence the rate of conception after myomectomy could be helpful for the decision. These factors are unknown at present (Vercellini et al., 1998).
Since 1989, we have set up continuous assessment with prospective collection of data concerning the patients operated in our department for myomectomy by laparoscopy (LM). The fertility results after LM for infertility were published previously (Dubuisson et al., 2000). These data were analyzed in order to identify the prognostic factors affecting fertility after myomectomy carried out in a context of infertility, and to clarify the indications for myomectomy for infertility.The prognositic factors were recently published ( Fauconnier et al,2000) ). The study included all infertile patients presenting a sub serous or intramural myoma measuring 20 mm or more and operated by LM between March 1989 and July 1996. Infertility was defined as the absence of all conception (whatever the outcome or location of the pregnancy) after at least 12 months attempt at pregnancy. The technique for LM was described previously (Dubuisson et al., 1997). For each patient operated by LM in a context of infertility, a pre operative work-up was made systematically including study of ovarian function (monthly temperature curve; F.S.H., L.H. levels, oestradiol, prolactin); partner's semen analysis; transvaginal ultrasonography; examination of the uterine cavity by diagnostic hysteroscopy or hysterosalpingography. The following per operative data was systematically collected: characteristics of myomas; existence of adnexal adhesions; tube appearance, permeability, and aspect of the mucosa; existence and extent of endometriosis.. The variables used for analysis were the following : age, duration of infertility, primary or secondary infertility; existence of menometrorrhagia pre operatively, existence of uterine cavity deformation; total number of myomas ; type (pedunculated, sessile or intramural), size (greatest diameter) and location (anterior, posterior or at the fundus) of the largest myoma; existence of a male factor defined by the presence of spermogram anomalies according to WHO criteria (W.H.O., 1987); existence of an ovulatory factor (anovulatory cycles on the monthly temperature curve with low or normal FSH, or hyperprolactinemia, or polycystic ovary); existence of tubal pathology (hydrosalpinx, phimosis, proximal anomalies); tubo-ovarian adhesions (whatever the degree of severity); existence of endometriosis according to the revised American Fertility Society classification (A.F.S., 1985); existence of surgery for infertility associated with the myomectomy and if so, what type. In July 1997 a postal questionnaire was sent to the women in the study. If no reply was received contact was made by telephone after checking for any change of address. When a patient was lost to follow-up, the patient's doctor was contacted. The questionnaire included how long attempts at pregnancy had lasted, any attempts at medically assisted procreation (MAP) together with the result, and for each pregnancy obtained after the myomectomy, the date it began and how it was achieved (e.g.: in vitro fertilization (IVF), stimulation, insemination, spontaneous) and its outcome (e.g.: spontaneous miscarriage, therapeutic abortion, voluntary abortion, birth). When pregnancy had been ongoing, a questionnaire was sent to the attending obstetrician.
The analysis criteria selected was the occurrence of spontaneous intra-uterine pregnancy whatever the outcome of this pregnancy. Intra-uterine pregnancies obtained by IVF together with ectopic pregnancies (EP) were excluded from analysis and the patients who obtained these pregnancies were censored at the point where their pregnancy began. Pregnancies obtained by stimulation, and those obtained by insemination were included with the spontaneous pregnancies. The cumulative probabilities of conception and their confidence interval at 95% (95% CI) were estimated using the Kaplan-Meier method (Kaplan and Meier, 1958). Initially the influence of each of the various variables on post-operative fertility were analyzed separately using the Log-rank test (Mantel, 1966). In order to determine what the independent prognostic factors for fertility are after LM we then carried out a multivariate analysis using Cox's regression model (Christensen, 1987; Cox, 1972). Variables with a significance threshold below 0.15 by univariate analysis were included in the regression. For the selection of the final model we used backward stepwise elimination method. Adjusted conception rate ratios (RR) and their 95% confidence intervals (95% CI) were calculated using the regression coefficients and their standart deviation of each variable of the final model. During the final phase of analysis we attempted to test the existence of interaction between the posterior nature of myomas and certain other variables in order to test the hypothesis that adnexal adhesions have an effect on fertility after myomectomy (Dubuisson et al., 1998; Tulandi et al., 1993; Ugur et al., 1996).
Ninety one patients were included in the study. Among the 91 operated patients, 81 (89%) of them were finally assessed for reproductive outcome. Seven patients were excluded from the follow-up: three because they no longer desired pregnancy after the LM; four because they had no chance of spontaneous conception (two cases of untreated bilateral tubal obstruction; one case of azoospermia; one case of anovulation with high F.S.H. levels). Three patients had no follow-up after myomectomy.
During the follow-up period 43 patients became pregnant after myomectomy. After exclusion of pregnancies obtained by IVF (n = 8) and EP (n = 1), the number of pregnancies taken into account in our analysis was 34, of which 7 were obtained by stimulation and 2 by intra-uterine insemination. The median duration of follow-up for patients who did not conceive was 23 months and the median time lapse prior to conception was 8 months. The cumulative probability of spontaneous intra-uterine conception was 34% (95% CI: 22 - 45) at 1 year, and 44% (95% CI: 32 - 56) at 2 years. The variables included in the regression model are those for which the threshold of significance was less than 0.15 at univariate analysis: the duration of infertility; existence of menometrorrhagia, the type and location of the largest myoma and the existence of an associated tubal pathology, male, or ovulatory factor. Fertility was higher in women who suffered from menometrorrhagia, due to their myoma(s), was lower in women whose largest myoma was intramural, in a posterior location, and who had an associated tubal pathology, male, or ovulatory factor. The association observed at univariate analysis between the duration of infertility and post-operative fertility disappeared once the other variables were taken into account, in particular the male and ovulatory factors. There was indeed an association between these two factors and the duration of infertility (p = 0.007 and p = 0.13 respectively).
If the largest myoma was intramural, this fact had an adverse effect on post-operative fertility only when it was in a posterior location (table 5), and the test for interaction was close to the threshold for significance (p = 0.06). Similar although non significant results were observed for the uterine suture, tubo ovarian adhesions before myomectomy and tubal pathology factor variables.
Post-operative fertility after myomectomy for infertility was lower when there was a posterior or intramural myoma, or one which had been sutured. It was higher after ablation of myomas responsible for menometrorrhagia. When there was an associated male factor, tubal pathology or ovulatory factor post-operative fertility was lower.
There are several points which are open to criticism in our study, however. (1) This was an observational study, and for certain patients pre operative assessment was not complete: for 5 patients the male factor was not investigated, and for these cases we considered that the results for the male factor were normal; the uterine cavity was not assessed pre operatively in 6 cases. (2) Data on postoperative fertility were collected retrospectively with a postal questionnaire (3) Certain characteristics of the myomas were not taken into account in our study: when the myoma is located on the uterine horns this could play a part in the infertility associated with the myoma (Ben-Ami et al., 1993; Gardner and Shaw, 1989; Kessel et al., 1988) ; the pre operative volume of the uterus could also play a part (Sudik et al., 1996). (4) Some of the factors studied were closely associated with each other: intramural nature and the need for uterine suture; the duration of infertility and associated infertility factors; the presence of tubo-ovarian adhesions and the existence of tubal pathology. All this needs to be taken into account when interpreting the results of the regression model.
The fact that the cumulative probability of conception after myomectomy was lower in the presence of a posterior myoma, an intramural myoma and/or uterine suture indirectly suggests the hypothesis that adhesions are responsible for the lower post-operative fertility. These factors were previously recognized as risk factors for adhesions after myomectomy: posteriors myomas (Dubuisson et al., 1998; Tulandi et al., 1993; Ugur et al., 1996) and/or intramural myomas (Ugur et al., 1996); the use of uterine sutures (Dubuisson et al., 1998; Elkins et al., 1987; Nezhat et al., 1991). The reduced fertility is secondary to involvement of the adnexa in adhesions associated with the scarring due to posterior myomectomy. Indirect support of this can be found in the analysis of the interaction between the posterior and intramural nature of the myoma, which shows that myomectomy of intramural myomas is only responsible for reduced fertility in case of posterior location. Several authors have postulated that adhesions might adversely affect fertility after myomectomy (Berkeley et al., 1983; Dubuisson et al., 1998; Gehlbach et al., 1993; Tulandi et al., 1993; Ugur et al., 1996). Our study provide indirect support of this in showing the role played for post operative fertility, by factors known to increase the adhesion rate.
The relationship between menometrorrhagia associated with a myoma and fertility after myomectomy needs to be interpreted with caution, given the low number of cases on which this result is based. Moreover, this relationship has never been found in other studies addressing fertility after myomectomy (Berkeley et al., 1983; Gehlbach et al., 1993; Smith and Uhlir, 1990; Starks, 1988; Sudik et al., 1996). However, one other author (Buttram and Reiter, 1981) did find a link between infertility and the existence of meno-metrorraghia (50% for patients suffering from infertility compared with 15 % of those without infertility). If this result is confirmed by other studies, it suggests that myomas responsible for menometrorrhagia are also responsible for infertility. Several physiopathological mechanisms can be suggested as an explanation for both infertility and meno-metrorraghia: (1) The existence of endometrial changes associated with the myomas could play a role (Deligdish and Loanthal, 1970; Hunt and Wallach, 1974; Stevenson, 1964), but these are related with sub mucosal myomas (Deligdish and Loanthal, 1970), which was not the case for the myomas in our sample. In fact, we did not look at the existence of this kind of endometrial change. (2) Deformation of the uterine cavity is often quoted to explain the association between infertility and the presence of a myoma (Buttram and Reiter, 1981; Deligdish and Loanthal, 1970; Hunt and Wallach, 1974; Ingersoll, 1963; Iosif and Akerlund, 1983; Sehgal and Haskins, 1960). In our study (table 3) like in others (Babaknia et al., 1978; Berkeley et al., 1983; Gehlbach et al., 1993; Smith and Uhlir, 1990; Starks, 1988; Sudik et al., 1996; Vercellini et al., 1998) we did not find that deformation of the uterine cavity had any influence on fertility after myomectomy. This is consistent with the results of two cross-matched studies (Eldar-Geva et al., 1998; Stovall et al., 1998) which showed that an intramural myoma could impair implantation even when there is no deformation of the uterine cavity. (3) The most likely explanation for the relation between menometrorrhagia (when there is no deformation to the cavity) and infertility is vascular changes (ectasia of the sub mucosal venous plexus) associated with the presence of the myoma (Farrer-Brown et al., 1971) which could play a part in infertility. These vascular changes play an important role in the onset of meno-metrorraghia (Buttram and Reiter, 1981; Farrer-Brown et al., 1971).
Intramural myomas themselves are known to hamper fertility (Eldar-Geva et al., 1998; Stovall et al., 1998). However in this study, the fertility after myomectomy for intramural myomas was lower than that after other types of myomas. This result can be explained by the fact that these intramural myomas were associated with an increased risk of post-operative adhesions. Anterior myomas and those at the fundus were not associated with the occurrence of adnexal adhesions, and post-operative fertility was equivalent to that for other types of myomas.
An unexpected finding of this study is that postoperative fertility was not affected by woman's age. Particularly the fertility only slightly decline in the group of women aging 40 years and over. In this group 6 women spontaneously conceived shortly after LM. It is well documented that fertility declines gradually over woman's reproductive lifespan (Federation CECOS et al., 1982; Tan et al., 1992). Two hypothesis could be drawn to explain this discrepancy: First of all, because of aging, some of the oldest women of the study could have been sent faster for IVF and thus prematurely censored. This could lead to some bias, but it is difficult to say if such a bias would explain our result. Moreover, because the incidence rate of myoma increases widely with age (Marshall et al., 1997), infertility, when really caused by myoma should, therefore, appear later in women's lifespan. In our study women were selected by having a myoma and being infertile and it is likely that these women would differ from the population of infertile women in general. For some of the women of our study the positive effect of the myomectomy might have compensated the negative effect of age in general. In many studies addressing fertility after myomectomy (Garcia and Tureck, 1984; Gatti et al., 1989; Gehlbach et al., 1993; Rosenfeld, 1986), no relationship was found between age and fertility. However, in one study which design was similar to ours it was found that aging was associated with the decrease of cumulative conception rate after myomectomy (Vercellini et al., 1999).
Although the presence of an associated infertility factor (tubal pathology, male, or ovulatory factor) was found to have an adverse effect on postoperative fertility, the design of our study does not allow any conclusions on the role of myomas itself as an infertility factor. Hence, it is difficult to say if myomectomy has an interest in these cases. Several authors consider that myomas could hinder the travel of the spermatozoids (Buttram and Reiter, 1981; Hunt and Wallach, 1974; Ingersoll, 1963; Iosif and Akerlund, 1983), but fertility after myomectomy was poor in cases with an associated male factor even after excluding one case with azoospermia. In cases with distal tuboplasty the rate of conception was low after myomectomy, but these rates differed little from those observed in series of distal plasty without myomectomy (Dubuisson et al., 1994; Hull, 1992). Would fertility after tuboplasty have been different if we had not treated the myoma? Myomectomy, could impair nevertheless the results of plasty due to the adhesions generated by the uterine scar. The fact that in cases of myomectomy for a posterior myoma the probability of conception after tubo-ovarian adhesiolysis or tubal surgery was lower than when the myoma was located elsewhere (table 5) provides indirect support to this.
When the workup for infertility reveals both a myoma and a major infertility factor (male, tubal or ovulatory), it is difficult to say if the myoma itself has much to do with the infertility. The indication for myomectomy should then be evoked in only three circumstances: (1) When the myoma is large, meaning there is a considerable risk of complications for future pregnancies (necrobiosis, miscarriage, threatened preterm delivery) (Exacoustos and Rosati, 1993). (2) When menometrorrhagia is associated with the myoma. (3) When IVF is indicated: if the myoma does not deform the cavity, the indications are not so clear (Fernandez, 1997), but the results of two controlled studies suggest that there is an advantage if myomectomy takes place in cases of even moderate sized intramural myomas, (Eldar-Geva et al., 1998; Stovall et al., 1998). The expected benefits must be weighed against the risk of uterine rupture after myomectomy both by laparotomy (Ozeren et al., 1997) and by laparoscopy (Dubuisson et al., 1995).