Human Reproduction vol.7 no.7 pp.973-977, 1992

The usefulness of endometrial biopsy for luteal phase evaluation in infertility

Juan Balasch1, Francisco Fibregues, Montserrat Creus and Juan A.Vanrell Department of Obstetrics and Gynaecology, Faculty of Medicine, Hospital Clinic i Provincial, 08036 Barcelona, Spain 'To whom correspondence should be addressed

To assess the usefulness of the late luteal phase endometrial biopsy in infertility, we evaluated a total of 1492 biopsies performed in 1055 patients. Of these women, 699 underwent one biopsy during spontaneous ovulatory cycles, 288 had two, 57 had three, nine had four, and five biopsies were done in two patients. As controls we included 45 fertile women who were requesting contraception. We analysed histological dating of the endometrium and its abnormality rates in first and successive biopsy specimens, as well as the association of the pregnancy outcome with the endometrial patterns and treatment for luteal phase deficiency (LPD). Our results show firstly that diagnosis of LPD in both infertile and fertile women represents only a chance event; secondly, histological endometrial adequacy or inadequacy in the cycle of conception or in previous cycles is not related to the outcome of pregnancy in infertile patients. Finally, treatment of LPD does not improve pregnancy outcome in infertile women. Thus, luteal phase evaluation by histological dating of the endometrium is not worthwhile. Key words: luteal phase deficiency/endometrial biopsy/infertility

Introduction Assessment of luteal function is considered an essential requirement in an infertility investigation. Luteal phase deficiency (LPD) is defined as a defect of progesterone output, by the corpus luteum, either in amount or duration, which results in inadequate stimulation of the endometrium for blastocyst implantation. In addition, to be considered as a factor responsible for infertility, LPD must be consistent and repetitive (Jones, 1975, 1976). On the basis of this definition, diagnosis of LPD has been traditionally made by premenstrual endometrial biopsy. Patients are diagnosed as having LPD (Jones, 1975, 1976) when endometrial biopsies in two different menstrual cycles show delayed histological maturation of > 2 days according to the dating criteria of Noyes et al. (1950). By using endometrial biopsy as the diagnostic technique, the reported incidences for LPD in infertility were between 3.5% and 20% (Balasch and Vanrell, 1987). The incidence is higher (23-60%) when only patients with repeated abortions are considered (Balasch and Vanrell, 1987). © Oxford University Press

However, the significance of endometrial LPD in fertility remains controversial. Thus, while diagnosis and treatment of the defect among patients with repeated abortion may improve the results of pregnancy (Balasch et al., 1986; Davidson et al., 1987), the implication in the literature that LPD must be identified and treated in order to improve the incidence and outcome of pregnancy in infertile patients has been challenged (Di Paola et al, 1971; Balasch et al., 1986; Wentz et al., 1990). In order to define the utility of endometrial biopsies in our own practice and to understand better the role of LPD in infertility, we carried out a retrospective analysis of the records of our infertile population.

Materials and methods All endometrial biopsies performed during spontaneous ovulatory cycles in infertile patients between 1975 and 1989 at the Infertility Unit of the Hospital Clinic i Provincial of Barcelona were identified by systematic review of the pathology and clinic records. Endometrial biopsies performed in patients with recurrent spontaneous fetal wastage and those carried out in infertile patients without basal body temperature records were not included in the study. As previously reported (Balasch et al., 1985), endometrial biopsies were performed in the late secretory phase (within 1 - 3 days of menstruation and at least 10 days after the nadir of basal body temperature). Endometrial samples were taken from high on the anterior and posterior walls of the uterine fundus and all of them were evaluated by the same expert gynaecological pathologist according to the histopathological criteria of Noyes et al. (1950). A biopsy specimen was considered abnormal when there was a delayed histological maturation of > 2 days. As controls, we included 45 fertile (mean parity 1.9, range 1 - 3 ) women (mean age 32 years, range 26-36 years) who were requesting contraception; 39 of them underwent one endometrial biopsy and six had two. The following items were analysed: (i) the abnormality rates in first and successive biopsy specimens in our infertile population and among the control group of fertile women; (ii) the pregnancy outcome according to the endometrial biopsy results; (iii) the association between endometrial histology in the cycle of conception and pregnancy outcome; (iv) the pregnancy outcome in 19 infertile patients achieving pregnancy while on treatment for LPD (treated group) and in those patients who became pregnant prior to therapy or after treatment was discontinued (n = 36). Patients were treated with vaginal progesterone, dehydrogesterone or clomiphene citrate plus human chorionic 973

J.Balasch et at.

gonadotrophin administered as previously reported (Balasch et al. 1982, 1983). Statistical analysis was performed using probability calculations, chi-square analysis with Yates' correction when necessary, or the Cochran Q-test for the significance of changes. Results A total of 1492 biopsies performed in 1055 infertile patients were included in the present study. Of these women, 699 underwent one biopsy during spontaneous ovulatory cycles, 288 had two, 57 had three, nine had four, and five biopsies were done in two patients. The mean age of patients included in the study was 29.8 years (range 18-44 years) and 126 of them (12%) had at least one child. The patients' characteristics are summarized in Table I.

Table I. Patients' characteristics No. of biopsies

1492

No. of patients One biopsy Two biopsies Three biopsies Four biopsies Five biopsies

1055 699 (66%) 288 (27%) 57 ( 5.5%) 9 ( 0.4%) 2 ( 0.5%)

Gravida 0

929 (88%)

Para a 1

126 (12%) 126 (12%)

Table D . Biopsy findings" in the 1055 patients and result of Cochran Q-test in patients with three or more biopsies'' Results of biopsies 1st

Number of patients 2nd

3rd

4th

5th 570 (54%) 110 (10.4%) 102 ( 9%) 92 (87%) 74 ( 7%)c 20 i 1.9%) 14 ( 1.3%) C 11 ( 1%) 0.6%) 7 i 7 i 0.6%) 7 ( 0.6%/ 7 ( 0.6%) 2 ( 0.2%) 2 ( 0.2%) 2 ( 0.2%/ ( 0.1%) ( 0.1%) ( 0.1%) ( 0.1%) I ( 0.1%) I ( 0.1%) 0.l%) c 0.1%) 0.1%)

Tuberculous endometritis

19 ( 1.8%)

*+ = In phase endometria, — = out of phase endometna b Q = 4.76; P = NS. ^ 5 patients fulfilling the two-biopsy criteria for diagnosis of luteal phase deficiency

974

The biopsy findings are summarized in Table II. Tuberculous endometritis was found in 19 patients (1.8%) and in all of these genital tuberculosis was suspected by hysterosalpingography. Excluding these 19 patients, 317 of 1036 endometria] biopsy specimens were out of phase on initial sampling (30.6%), and on rebiopsy, 95 of 207 patients (45.8%) had persistently out of phase specimens, thus fulfilling the two-biopsy criteria for diagnosis of LPD. To assess whether the 95 cases of LPD (10.3% of 926 women, including the 719 patients with in phase endometria on initial sampling and the 207 rebiopsied patients after an out of phase first endometrial specimen) were more than would be expected solely on the basis of chance, probability calculations were used. With a baseline rate of 317 of 1036 endometrial biopsies being out of phase, the chance of a randomly selected sample being out of phase was 317 of 1036. If the results of two consecutive endometrial specimens were random, then we would have expected two consecutive endometrial biopsies to be out of phase by chance with a probability: (317/1036) x (317/1036) = 0.306 X 0.306 = 0.0936. Of 926 women we would have expected 926 x 0.0936 or 86.7 women with LPD by chance. We actually observed 95 patients with LPD. These figures (expected and observed) were similar by chi-square test (X2 = 0.39, P = NS). Of the 926 women, there were 119 aged >35 years and 14 of these 119 patients (11.7) had LPD. This figure was similar to the 10% incidence found among women aged 2 days in eight of the 45 (17.8%) fertile women. Six of these eight women underwent a second biopsy in a later cycle and two of therebiopsieswere out of phase. Thus we would have expected two consecutive biopsies to be out of phase by chance with a probability: (8/45) x (8/45) = 0.178 X 0.178 = 0.0316. Of 43 women we would have expected 43 x 0.0316 or 1.36 women with LPD by chance. We actually observed two patients with LPD (4.7%). These figures of expected and observed were similar by chi-square test (x2 = 0.04, P = NS). Furthermore, there was no significant difference between the control and infertile groups (x2 = 0.88, P = NS). Of the 335 spontaneous pregnancies observed among the patients studied, 283 (82.3%) were viable pregnancies, 52 ended in spontaneous abortion (15.1%), there were eight (2.3%) ectopic pregnancies and one (0.3%) hydatidiform mole. The distribution of pregnancies (viable or non-viable) according to the endometrial biopsy results is shown in Table in. No association between histological findings and pregnancy outcome was found by chi-square test. An endometrial biopsy was performed in a pregnancy cycle in 27 patients (1.7%). In all instances the endometrial specimens were noted to be fundal samples that were clearly progestational, but abnormal secretory phases were detected in six patients

Endometrial biopsy in infertility

Table III. Pregnancy outcome according to the previous endometrial biopsy results in 335 patients1 Pregnancy outcome Term Abortion Total

Results of biopsiesb

+

-

++

155 27 182

31 2 33

28 5 33

+4 3 7

- -

- +

+++

19 3 22

24 4 28

6 2 8

h

7 3 10

1 1 2

H

- + -\-

+ - 4•

++ -

- - ++

+ ++ -

Total

0 1 1

3 0 3

2 0 2

2 0 2

0 1 1

1 0 1

283 52 335

V (with 14 degrees of freedom) = 22.724, P = NS. b + = In phase, - = out of phase.

(22.2%). In no case was gestational hyperplasia reported by the pathologist. The outcome of these 27 pregnancies was as follows: 19 (71%) normal deliveries at term, seven first-trimester abortions (26%), and one hydatidiform mole (3%). The 26% abortion rate in this subgroup of patients was not significantly different from the 15% incidence in the total study group (x2 = 2.29, P = NS). There was no association between endometrial histology in the cycle of conception and the outcome of pregnancy (Table IV). Table V shows the outcome of pregnancy in treated and untreated infertile patients who were diagnosed as having LPD by the two-biopsy criteria. No difference in the pregnancy outcome was found between treated and untreated groups in the cycle of conception. Discussion Premenstrual endometrial biopsy is usually employed both as a bioassay of the steroidogenic function of the corpus luteum and as a means to assess normality of the implantation site in the infertile patient (Jones, 1975, 1976); however, the endometrial LPD is associated with normal hormonal levels in the great majority of infertile patients, thus indicating an intrinsic lack of responsiveness of the target tissue (Balasch and Vanrell, 1987). It has been suggested that women aged ^35 years may have a higher incidence of LPD (Wentz, 1988). However, in our study, age had no significant effect on the biopsy result. The morphological criteria used to assess the histological dating of endometrium have been described in detail by Noyes et al. (1950). They found that the histological dating of the endometrial pattern in relation to the duration of the luteal function was reproducible within a 2 day interval. This is still considered to be the most complete description of the continued, predictable development of the luteal phase endometrium (Wentz, 1988). Presumption of ovulation may be easily made in most patients, but despite the rigid criteria used to evaluate the biopsy, the correct interpretation may be fairly difficult because of subtle daily changes, variations in fixation and quantity of biopsy material (Davidson el al., 1987). Furthermore, in many institutions the endometrial biopsy is interpreted by general pathologists, many of whom may lack experience or interest in gynaecological pathology (Davidson et al., 1987; Wentz, 1988). Therefore, this may be the reason for a missed diagnosis in a fair number of patients (Davidson et al., 1987). In our hospital, all biopsies are read by the same expert gynaecological pathologist, which reduces inter-evaluator variation and adds to the accuracy of the diagnoses (Gibson et al., 1991).

Table FV. Pregnancy outcome according to endometrial histology in the cycle of conception" Pregnancy outcome Term

Endometrial histology In phaseb

Out of phase

15(79%)

4(21%)

5 (72%)

2 (28%)

Spontaneous abortion

V = 0.16: P = NS. b Orte hydatidiform mole excluded (endometrial histology: in phase).

Table V. Pregnancy outcome in patients treated for luteal phase deficiency (LPD) and untreated patients3 Treatment for LPD

Yes Noc

b

Pregnancy outcome Term

Spontaneous abortion

14 (78%) 27 (77%)

4 (22%) 8 (23%)

" x 2 = 0.003: P = NS. ''One hydatidiform mole excluded. c One ectopic pregnancy excluded

Nine years after the original description of the criteria involved in dating endometrial biopsies, Noyes (1959) critically and extensively examined the problem of the under-developed secretory endometrium, particularly as to its incidence, the difficulties encountered in its diagnosis and treatment and its significance for the implantation and development of the ovum. On the basis of clinical and experimental data, he concluded that the significance of the under-developed secretory endometrium had been overrated, and that endocrine treatment of infertile and aborting patients for LPD should be considered as empirical rather than rational therapy (Noyes, 1959). Curiously, however, the value of biopsy in diagnosis of LPD has stood the test of time. One of the most intriguing facts in infertility is whether one cyclic event may be representative of all patient cycles. It is generally accepted that to be of clinical significance, LPD must be present in at least two separate cycles (Jones, 1975, 1976). Some authors claim that the incidence of this defect and the response to treatment are frequent enough to make mandatory the routine screening of infertility patients by timed endometrial biopsy (Soules et al., 1977; Rosenberg et al., 1980; Daly et al., 1983), but others disagree. The poor prognostic significance of the initial endometrial biopsy has been emphasized in several reports (Stevenson, 1965; Murthy et al., 1970; Driessen et al., 975

J.Balasch et al.

1980). In patients diagnosed as having LPD by the two-biopsy criteria, Di Paola et al. (1971) found no difference in the incidence and outcome of pregnancy between couples in whom LPD was the sole infertility factor and those with other causes of infertility. They concluded that this entity does not represent an infertility factor. Similarly, Davidson et al. (1987) reported that endometrial biopsies may be useful only if performed in cases of habitual abortion or ovulation induction with clomiphene citrate. More recently, Wentz et al. (1990) analysed the impact of LPD in an infertile population of 137 women and concluded that there was no difference observed in fecundity between patients diagnosed with or without LPD. Our findings confirm and extend previous studies suggesting the lack of association between LPD, as diagnosed by endometrial biopsy, and infertility. We show that the diagnosis of LPD represents only a chance event both on our infertile population and among a control group of fertile women. This latter aspect agrees with our previous report revealing a similar incidence of LPD among a smaller number of infertile and fertile women (Balasch et al., 1986). In addition, analysis of successive endometria] biopsies in infertile patients, with three or more biopsy specimens from separate cycles, further emphasizes that the results observed are random. The lack of significance of LPD for fertility is supported by our data showing that there is no association between the outcome of pregnancy and the previous endometrial biopsy results, the endometrial histology during the conception cycle, or the treatment for LPD. It should be noted that seven out of 10 patients achieved term pregnancy without treatment after being diagnosed as having LPD by three endometrial biopsies in different cycles (Table HI). The effects of the biopsy during the cycle of conception on the pregnancy are controversial. Thus, Karow et al. (1971) reported an average abortion rate of 6.8% in their collective review of 152 endometrial biopsies done in early pregnancy; and based on the Loeb reaction observed in animal experimental models, they suggested a possible therapeutic benefit of luteal phase endometrial biopsy. These authors postulate that mechanical stimulation of the cervix and traumatization of the uterus in some way provides an endocrine stimulus of the corpus luteum and thus leads to a more favourable environment for the developing ovum. However, analysis of the data in the literature reveals a 14.6% (range 0.0—66.6%) incidence of fetal wastage among pregnancies with inadvertant endometrial biopsy in the cycle of conception (Balasch et al., 1984). This incidence is similar to the spontaneous abortion rate of the general population and argues against the claimed benefits of endometria] biopsy in the conception cycle. In agreement with Rosenfeld and Garcia (1975), patients with out of phase secretory endometria in the cycle of conception who deliver normal, live, term infants further illustrate the difficulties in making the diagnosis of the LPD through histological assessment of the endometrium. Our data concur with those of Rosenfeld and Garcia (1975) in that gestational hyperplasia was not diagnosed in any instance. Although some authors state that treatment of infertile patients suffering from LPD results in improved fertility (Soules et al., 1977; Rosenberg et al., 1980; Daly et al., 1983), others dispute these findings (Balasch et al., 1986; Davidson et al., 1987; Wentz et al., 1990) and the claims 976

that treatment is effective have not been substantiated by controlled trials. There is no doubt that a normal histological, endocrinological and biochemical milieu is a prerequisite to normal nidation (Wentz, 1988). According to our results, however, the dated endometrial biopsy is but a gross indicator of the milieu. Thus, the usefulness of the luteal phase evaluation in infertility, by histological dating of the endometrium, should be questioned and we clearly need other markers of adequate endometria] differentiation. Recently, there have been attempts to improve endometria] dating by more modern approaches (Li et al., 1988, 1992; Li and Cooke, 1991). In summary, these studies suggest that it is unlikely that ultrasonographic measurement of endometrial thickness can replace histological examination of the endometrium in the evaluation of the luteal phase (Li et al., 1992). However, the precision of endometria] dating may be improved if chronological dating is based on the luteinizing hormone surge or the time of follicle rupture (determined by serial ultrasonography), and if histological dating is based on morphometry, a quantitative and objective histological technique (Li et al., 1988; Li and Cooke, 1991).

Acknowledgement We gratefully acknowledge the assistance of Dr J.Aznar in performing the statistical evaluation of the results. References BalaschJ. and Vanrell.J.A. (1987) Corpus luteum insufficiency and fertility: a matter of controversy. Hum. Reprod., 2, 557—567. Balasch.J., Vanrell.J.A., M^rquez.M., Burzaco.I. and GonzilezMerloJ. (1982) Dehydrogesterone versus vaginal progesterone in the treatment of the endometrial luteal phase deficiency. Fertil. Steril., 37, 751-754. Balasch.J., Vanrell,J.A., Mdrquez.M. and GonziJez-Merlo.J. (1983) Dehydrogesterone treatment of endometrial luteal phase deficiency after ovulation induced by clomiphene citrate and human chorionic gonadotropin. Fertil. Steril., 40, 469-471. Balasch.J., Vanrell.J.A., Ma>quez,M. and Gonzilez-Merlo.J. (1984) Endometrial biopsy inadvertently taken in the cycle of conception. Int. J. Gynaecol. Obstet., 22, 9 5 - 9 9 . BalaschJ., Vanrell.J.A., Creus.M., Mfirquez.M. and Gonzalez-MerloJ. (1985) The endometrial biopsy for diagnosis of luteal phase deficiency. Fertil. Steril., 44, 699-701. BalaschJ., Creus.M., Mdrquez.M., Burzaco.I. and Vanrell.J.A. (1986) The significance of luteal phase deficiency on fertility: a diagnostic and therapeutic approach. Hum. Reprod., 1, 145 — 147. Daly.D.C, Walters.C.A., Soto-Albors.C.E. and Riddick.D.H. (1983) Endometrial biopsy during treatment of luteal phase defects is predictive of therapeutic outcome. Fertil. Steril., 40, 305-310. Davidson.B.J., Thrasher.T.V. and Seraj.l.M. (1987) An analysis of endometnal biopsies performed for infertility. Fertil. Steril., 48, 770-774. Di Paola.G.R., M^ndez Ribas.J.M. and Arrighi.L.A. (1971) Critical study of theretardedprogestational phase. Int. J. Fertil., 16, 189-194. Driessen.F., Holwerda.P.J., Putte.S.C. and Kremer.J. (1980) The significance of dating an endometrial biopsy for the prognosis of the infertile couple. Int. J. Fertil., 25, 112-116. Gibson,M.. Badger,G.J., Byrn,F., Lee,K.R., Korson.R. and Trainer.T.D. (1991) Error in histologic dating of secretory

Endometria] biopsy in infertility

endometrium: variance component analysis. Fertil. Steril., 56, 242-247. Jones,G.S. (1975) Luteal phase defects. In Behrman.S.J. and Kistner.R.W. (eds), Progress in Infertility, 2nd edn. Little, Brown and Co., Boston, p. 299-324. Jones.G.S. (1976) The luteal phase defect. Fertil. Steril., 27, 351-356. Karow,W.G., Gentry.W.C, Skeels,R.F. and Payne,S.A. (1971) Endometrial biopsy in the luteal phase of the cycle of conception. Fertil. Steril., 22, 482-495. Li,T.C. and CookeJ.D. (1991) Evaluation of the luteal phase. Hum. Reprod., 6, 484-499. Li,T.C, Rogers,A.W., Dockery,P., Lenton.E.A. and Cooke.I.D. (1988) A new method of histologic dating of human endometrium in the luteal phase. Fertil. Steril., 50, 52-60. Li,T.C, Nutlall,L., Klentzeris.L. and Cooke.I.D. (1992) How well does ultrasonographic measurement of endometrial thickness predict the results of histological dating? Hum. Reprod., 7, 1-5. Murthy,Y.S., Arronet,G.H. and Parekh.M.C. (1970) Luteal phase inadequacy. Its significance in infertility. Obstet. Gynecol., 36, 758-761. Noyes,R.W. (1959) The underdeveloped secretory endometrium. Am. J. Obstet. Gynecol., 77, 929-945. Noyes,R.W., Hertig,A.T. and Rock,J. (1950) Dating the endometrial biopsy. Fertil. Steril., 1, 3 - 2 5 . Rosenberg,S.M., Luciano.A.A. and Riddick,D.H. (1980) The luteal phase defect: The relative frequency of, and encouraging response to, treatment with vaginal progesterone. Fertil. Steril., 34, 17—20. Rosenfeld.D.L. and Garcia,C.R. (1975) Endometrial biopsy in the cycle of conception. Fertil. Steril., 26, 1088-1093. Soules.M.R., Wiebe.R.H., Aksel,S. and Hammond.C.B. (1977) The diagnosis and therapy of luteal phase deficiency. Fertil. Steril., 28, 1003-1037. Stevenson,C.S. (1965) The endometrium in infertile women: prognostic significance of the initial study biopsy. Fertil. Steril., 16, 208—222. Wentz.A.C. (1988) Luteal phase inadequacy. In Behrman.S.J., Kistner,R.W. and Patton.G.W. (eds), Progress in Infertility, 3rd edn. Little, Brown and Co., Boston, pp. 405-462. Wentz,A.C, Kossoy.L.R. and Parker.R.A. (1990) The impact of luteal phase inadequacy in an infertile population. Am. J. Obstet. Gynecol., 162, 937-945. Received on March 18, 1992; accepted on May 8, 1992

977