Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
Prediction of Implantation by the Sonographic Appearance of the Endometrium During Controlled Ovarian Stimulation for in Vitro Fertilization (IVF) YAEL GONEN 1'2 and ROBERT F. CASPER 1'3
Submitted: November 23, 1989 Accepted: February 20, 1990
INTRODUCTION
The texture and the thickness o f the endometrium as assessed by transvaginal sonography were prospectively evaluated in 123 patients undergoing I V F treatment. Three different types o f endometrial patterns could be distinguished: (A) an entirely homogeneous, hyperechogenic endometrium; (B) an intermediate type characterized by the same reflectivity o f ultrasound as the myornetrium, with a nonprominent or absent central echogenic line; and (C) a multilayered endometrium consisting o f prominent outer and midline hyperechogenic lines and inner hypoechogenic regions. On the day before oocyte retrieval, endometrial thickness was significantly greater in the group o f patients who achieved pregnancy than in the group who did not (8.7 +--0.4 vs 7.5 +- 0.2 ram, respectively; P < 0.01) and significantly more patients had multilayered, pattern C, endometrium (75% in pregnant women vs 42.4% in nonpregnant women; P < 0.01). N o pregnancy occurred when the endometrial thickness was less than 6 mm. When type C endometrium >-6 mm thick was seen, the pregnancy rate per embryo transfer was 39%. When type A or B endometrial pattern was seen, the negative predictive value f o r the occurrence o f pregnancy was 90.5%. Our results suggest that transvaginal sonographic evaluation o f endometrial texture and thickness may be an indicator o f the likelihood o f achieving pregnancy.
The introduction of the vaginal ultrasound transducer, with its high-frequency, high-resolution capability, represents a major advance in gynecologic sonography. Transvaginal ultrasound has focused mainly on measurement of follicular diameter for cycle monitoring, on its use as an improved modality for oocyte retrieval for in vitro fertilization (IVF), and on the early detection of intrauterine pregnancy by visualization of the gestational sac. Recently, we (1) and others (2) have shown that endometrial thickness and the amount of endometrial growth during controlled ovarian stimulation for IVF are significantly greater in conception cycles compared to nonconception cycles, a finding which may have potential predictive value with regard to implantation. In the present study, we attempt to extend these findings by examining both the thickness and the sonographic pattern of the endometrium in patients undergoing IVF. Different patterns of endometrial response in natural and stimulated cycles have been described (3-5) but a relationship between endometrial thickness and reflectivity to ultrasound has not been determined. Recently, Welker and coworkers (6) demonstrated that specific patterns of endometrial development can be distinguished by transvaginal ultrasound during IVF cycles and may be predictive of implantation. The objective of the present study was to determine, using vaginal sonography, whether a particular combination of endometrial pattern and growth during ovarian hyperstimulation for IVF could predict the subsequent occurrence of pregnancy.
KEY WORDS: transvaginal sonography; endometrial texture; endometrial thickness; in vitro fertilization.
i Division of Reproductive Science, Department of Obstetrics and Gynecology, The University of Toronto, Toronto, Canada. 2 Present address: Department of Obstetrics and Gynecology, Carmel Hospital, Haifa, Israel. 3 To whom correspondence should be addressed at 6-240 EN, Toronto General Hospital, 200 Elizabeth Street, Toronto, Ontario, Canada M5G 2C4.
0740-7769/90/0600-0146506.00/0 9 1990 Plenum Publishing Corporation
146
ENDOMETRIAL PATTERN AND THICKNESS DURING IVF
MATERIALS AND METHODS Subjects Endometrial thickness and pattern were prospectively studied during induction of ovulation in 123 consecutive patients undergoing treatment in the University of Toronto IVF program. The mean (-+SE) age of the women was 32.9 -+ 0.31 years, with a range from 25 to 39 years. Most of the patients (69.9%) had tubal infertility, 9.8% had endometriosis, and 8.9% were suffering from longstanding infertility of unknown etiology. Male factor was involved as the cause of infertility either as the sole problem or combined with tubal factor or endometriosis in 11.4% of patients. All patients included in the study underwent standardized ovarian follicular stimulation with clomiphene citrate (Serophene, Serono, Randolph, MA), 100 mg/day, cycle days 5-9, and human menopausal gonadotropins (hMG; Pergonal, Serono), 75 to 150 IU im starting on cycle day 5 or 6. Human chorionic gonadotropin (hCG; Profasi, Serono), 5000 IU im, was routinely used to trigger the final stage of follicular maturation. Oocyte retrieval and fertilization were performed using standard techniques in each patient without regard to the endometrial findings. Embryo transfer occurred 48 hr after retrieval in all patients. No luteal-phase support was used. Cycle Monitoring The response to treatment was monitored daily starting on day 9 of the menstrual cycle by daily determination of serum levels of estradiol (E2) and luteinizing hormone (LH) by radioimmunoassay (RIA) as previously described (7,8). Transvaginal ultrasound measurements of follicular diameter and endometrial thickness were recorded daily, starting on cycle day 10. Endometrial thickness was measured on the "frozen" ultrasound image obtained with a 7-MHz vaginal transducer (Bruel & Kjaer, Type 8538, Naerum, Denmark). The maximum thickness of endometrium on both sides of the midline was measured in the plane through the central longitudinal axis of the uterine body (Fig. I). Endometrial Pattern and Texture
Pictures of the endometrium were taken on the day after administration of hCG (the day prior to
147
ovum retrieval) using a Mitsubishi video copy processor (Type P-60, Japan) and were stored for later evaluation of endometrial pattern. Near and far gains were constant for all measurements. Both investigators independently assessed endometrial pattern in all patients with subjective agreement. Endometrial pattern was determined by comparing the reflectivity or gray-scale appearance of the endometrium to that of the adjacent myometrium. The changes in reflectivity varied from black or darker than the myometrium (hypoechogenic), to equal to the appearance of the myometrium (isoechogenic), to brighter than the myometrium (hyperechogenic). Three endometrial ultrasound patterns were distinguished. (A) An entirely homogeneous, hyperechogenic endometrium was characterized by increased reflectivity compared to the myometrium and consequently appeared brighter in its gray-scale appearance. The central echogenic line was not visualized (Fig. la). (B) An intermediate, isoechogenic pattern was characterized by the same reflectivity compared to the surrounding myometrium, with a nonprominent or absent central echogenic line (Fig. lb). (C) A multilayered "triple-line" endometrium consisted of prominent outer and central hyperechogenic lines and inner hypoechogenic or black regions (Fig. lc). Statistical Evaluation
The data obtained were compared by Student's t test for unpaired results and the chi-square test. P values 1.5 cm 6.2 - 0.3 5.9 -+ 0.3 No. oocytes per retrieval 5.4 --+ 0.4 5.5 --- 0.5 No. embryos per transfer 3.0 -+ 0.3 2.6 --- 0.4 a
Not significant.
NS a NS NS NS NS NS NS
Thirty-one of the 99 nonpregnant patients did not have an embryo transfer due to lack of fertilization, and in 3 others, no eggs were retrieved. In this group, 47% (16/34) displayed a pattern A endometrium. Significantly more patients with patterns A and B endometrium failed to achieve embryo transfer (23/63; 36.5%) than patients with pattern C (11/60; 18.3%; P < 0.05) (Table I). There was no difference in endometrial thickness between the group of patients with lack of fertilization and those who did not conceive after embryo transfer (7.3 -+ 0.4 and 7.7 - 0.2 mm, respectively).
DISCUSSION We found, using transvaginal sonography, that the presence of pattern A or B endometrium on the day after hCG administration (day prior to oocyte recovery) was associated with a negative predictive value for pregnancy of 90.5%. In addition, in over 60 IVF pregnancies obtained by us in the past year, none occurred in any patient found to have an endometrial thickness below 6 mm on the day after hCG. In contrast, the presence of pattern C endometrium of 6-mm thickness or greater on the day after hCG administration was associated with more than a 39% clinical pregnancy rate per embryo transfer in our IVF program. These data confirm the results of Welker et al. (6), who also reported higher pregnancy rates in women with endometrium displaying a triple-line pattern (pattern C) on the day of oocyte retrieval than in those with a hyperechogenic endometrium (pattern A). With these findings to date, we are quite confident that we can determine those patients who have a very poor chance of conceiving in a specific IVF cycle prior to the oocyte recovery. Furthermore, our results demonstrate that the endometrial pattern and thickness on the day before oocyte recovery, if unfavorable, are predictive of a negative outcome. The quality and state of preparation of the endometrium are two of several factors which may influence the success or failure of in vitro fertilization and embryo transfer (9). By using ultrasound examination, it is now possible to determine routinely and accurately, the thickness and texture of the endometrium (9-11). The transvaginal, highfrequency transducer enables even more detailed visualization of small changes in endometrial pattern and growth during the cycle. The criteria we studied, namely, endometrial pattern and thickness,
Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
150
GONEN AND CASPER
100 -
J
ill PREGNANT [ ] NONPREGNANT
I
80,
I'-
50'
zi l l r.3 iz: i,u n
40'
20'
.2-.3
.4-.5
.6-.7
.8-.9
1.0-1.1 1.2-1.3 1.4-1.5
ENDOMETRIAL THICKNESS (CM) Fig. 2. The occurrence of clinical pregnancy related to endometrial thickness. Percentage, on the Y axis, refers to the percentage of each group [pregnant cycles (hatched bars) and nonpregnant cycles (open bars)] which occurred at each level of endometrial thickness.
appear to be useful in prediction of pregnancy occurrence, especially when an unfavorable endometrial picture is visualized. This unfavorable pattern consists of three factors: (i) a relatively hyperechogenic or isoechogenic texture compared to the myometrium, (ii) the absence of hyperechogenic lines separating the endometrium from the myometrium and the absence or faint presence of a central echogenic line, and (iii) a thickness of less than 6 mm. In contrast, a favorable endometrial pattern for the subsequent occurrence of pregnancy appears as multilayered endometrium consisting of prominent outer and midline hyperechogenic lines and inner hypoechogenic or black regions and an endometrial thickness of at least 6 mm. This favorable pattern has been described by Forrest et al. (12) as typical of the proliferative endometrium in natural cycles and characterized by these authors as the appearance of "triple lines" running longitudinally in the center of the uterus. The central echogenic line is produced by the endometrial canal but the histologic structures that give rise to the echogenic outer lines are not known. Forrest et al. (12) believe that these outer lines may represent the basal layer of the endometrium or the interface between the endometrium and the myometrium. They
also describe clearly the relatively hypoechogenic regions between the two outer lines and the central line, which probably represents the functional layer of the endometrium. As endometrial development progresses during the late proliferative phase, the functional layer thickens and the distance between the outer two lines and the central echogenic line increases. Although, in general, a favorable, type C pattern of endometrial reflectivity is associated with a prominent endometrial thickness, we have occasionally found that an adequate increase in the thickness of the endometrium may fail to occur with this pattern. In all cases where the endometrial thickness was found to be less than 6 mm on the day prior to oocyte recovery, regardless of endometrial pattern, clinical pregnancy subsequently failed to occur in our program. We have previously demonstrated that fertilization rate and endometrial thickness are not correlated, and that these two measurements appear to be independent of each other (1). In contrast, Smith et al. (5) showed, using retrospective analysis, that an endometrial thickness of less than 5 mm (with transabdominal sonography) was associated with either failure to recover oocytes or the recovery of
Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
151
ENDOMETRIAL PATTERN AND THICKNESS DURING IVF
TYPE "A"
20-
15 !i!i~!i!~!i!!!~!~i~i!i!i
i~i!~!~:i!i!i!~!i!i~iil
10 ::::::::.::::,
:::::::::,:::
......................
~:~:~:~:~:~:~:~:~:~:~:~:7 ,::::,::::.::: .................. ::::::::::::::::::::::
::::::::::::::::::::::: ........................ ::::.::::::::: ::::::::::::
::::::::::::
.....................
~:~:~.::~:~::::
::::::::::::::::::::::
::::::::::::::::::::::::
0.2-0.3
0.4-0.5
:::::::::::::.::::::
0.6-0.7
0.8-019
1.o-1.1
1,2-1.3
1.4-1.5
TYPE "B" 10
U~ IJ.I
f
8
:::+:::,::::.:
.-I
r >.. r
:~::~::::.:,::,:.:,
6
4
Ii
O :it
2 ::,::::.::::.: ::::::::::::::::::::::::
0 20"
0.4-0.5
0.6-0.7
iiil
0.8-0.9
1.0-1.1
1.2-1.3
1.4-1.5
TYPE "C" 15 84
1 0 84
5'
0.4-0.5
0,6-0.7
0.8-0.9
1.0-1.1
ENDOMETRIAL THICKNESS
1.2-1.3
(cm)
Fig. 3. The relationship among endometrial pattern, endometrial thickness, and the occurrence of clinical pregnancy. (a) Pregnancy cycles (hatched bars) and nonpregnancy cycles (open bars) demonstrating endometrial pattern A (hyperechogenic) on the day after hCG administration related to endometrial thickness on the X axis. No pregnancy occurred with endometrial pattern A and a thickness less than 10 mm. (b) Pregnancy and nonpregnancy cycles in patients with endometrial pattern B (intermediate) on the day after hCG administration related to endometrial thickness. No pregnancy occurred with endometrial pattern B and thickness less than 8 mm. (c) Pregnancy and nonpregnancy cycles in patients with endometrial pattern C (multilayered) on the day after hCG. No pregnancy occurred if endometrial thickness was less than 6 mm.
immature oocytes with a lower fertilization rate. In the present study, the mean endometrial thickness in patients with lack of fertilization did not differ from the mean endometrial thickness of patients
who achieved fertilization and did not conceive, consistent with the findings of our previous study (1). In contrast, poor endometrial patterns (A and B) were associated with twice the incidence of fail-
Journal o f in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
152
GONEN AND CASPER
ure to fertilize than pattern C (multilayered pattern). These data suggest that changes in endometrial pattern during IVF stimulation may also be useful as predictive parameters for fertilization. In keeping with the findings of our present study, Welker et al. (6), also using transvaginal ultrasound, reported that the pattern of the endometrium is a predictive parameter for conception and suggested cancellation of oocyte collection or embryo replacement when a homogenous, hyperechogenic endometrium is present. In contrast to our results, these investigators did not find a correlation between endometrial thickness and occurrence of pregnancy. Glissant et al. (2) reported that endometrial thickness was significantly greater in conception compared to nonconception cycles but concluded that it was not possible to predict the probability of pregnancy with the use of endometrial thickness. We agree that the positive predictive value of endometrial thickness is low. However, our experience indicates that utilization of the combination of endometrial pattern and endometrial thickness has great negative predictive value for the subsequent occurrence of pregnancy. We believe that the findings of the present study will be helpful in decision-making regarding oocyte recovery, embryo replacement, and cryopreservation. If pattern A or B endometrium is seen or if the endometrial thickness is less than 6 mm with any pattern on the day after hCG administration, consideration should be given to oocyte recovery with transfer of frozen/thawed embryos during a subsequent cycle when more favorable endometrium is present. A few of the questions which remain to be answered include the following. (i) Does an unfavorable pattern or thickness of the endometrium recur in the same patient? (ii) Does the method of ovarian hyperstimulation contribute to the pattern and thickness of the endometrium? For example, could an antiestrogenic effect of clomiphene citrate causing estrogen receptor depletion in the endometrium be the cause of the unfavorable endometrial patterns and texture seen in the present study? (iii) Do unfavorable endometrial patterns or thicknesses occur in natural cycles? (iv) Can an unfavorable endometrial pattern or thickness be improved by hormonal manipulation or alteration of stimulation protocol? Further studies will be needed to address these and other questions concerning the preliminary findings of this report.
ACKNOWLEDGMENTS This work was supported by a research grant from the Medical Research Council of Canada and by a Fellowship Training Grant to Dr. Y. Gonen from Wyeth Pharmaceutical Company, Toronto, Ontario, Canada This paper was presented in part at the Postgraduate Course on Advanced IVF at the American Fertility Society Meeting, San Francisco, California, November 12, 1989. REFERENCES 1. Gonen Y, Casper RF, Jacobson W, Blankier J: Endometrial thickness and growth during ovarian stimulation: A possible predictor of implantation in in vitro fertilization. Fertil Steril 1989;52:446--450 2. Glissant A, de Mouzon J, Frydman R: Ultrasound study of the endometrium during in vitro fertilization cycles. Fertil Steril 1985;44:786-790 3. Fleischer AC, Kalemeris GC, Entman SS: Sonographic depiction of the endometrium during normal cycles. J Ultrasound Med Biol 1986;12:271-277 4. Fleischer AC, Pittway DE, Beard LA, Thieme GA, Bundy AL, James AE Jr, Wentz AC: Sonographic depiction of endometrial changes occurring with ovulation induction. J Ultrasound Med 1984;3:341-346 5. Smith B, Porter R, Ahuja K, Craft I: Ultrasonic assessment of endometrial changes in stimulated cycles in an in vitro fertilization and embryo transfer program. J Vitro Fert Embryo Transfer 1984;1:233-238 6. Welker BG, Gembruch U, Diedrich K, A1-Hasani S, Krebs D: Transvaginal sonography of the endometrium during ovum pickup in stimulated cycles for in vitro fertilization. J Ultrasound Med 1989;8:549--53 7. Gonen Y, Casper RF: Does transient hyperprolactinemia during ovarian hyperstimulation interfere with conception or pregnancy outcome. Fertil Steril 1989;51:1007-1010 8. Casper RF, Armstrong DT, Brown SE, Daniel SAJ, Yuzpe A, Erskine H: In vitro fertilization: Determination of follicular maturation for timing of human chorionic gonadotropin administration. Fertil Steril 1987;47:345-349 9. de Mouzon J, Lefevre B, Frydman R, Belaisch-Allert JC, Guillet-Rosso F, Testart J: Factors affecting human in vitro fertilization: a multifactorial study. Fertil Steril 1985;43:892896 10. Brant T, Levy E, Grant T, Marut E, Leland J: Endometrial echo and its significance in female infertility. Radiology 1985;157:225-229 Fleisher A, Herbert CM, Sacks GA, Wentz A, Entman S, James A: Sonography of the endometrium during conception and nonconception cycles of in vitro fertilization and embryo transfer. Fertil Steril 1986;46:442-447 12. Forrest TS, Elyaderni MK, Muilenburg MI, Bewtra C, Kable WT, Sullivan P: Cyclic endometrial changes: US assessment with histologic correlation. Radiology 1988;167: 233-237
Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
Prediction of Implantation by the Sonographic Appearance of the Endometrium During Controlled Ovarian Stimulation for in Vitro Fertilization (IVF) YAEL GONEN 1'2 and ROBERT F. CASPER 1'3
Submitted: November 23, 1989 Accepted: February 20, 1990
INTRODUCTION
The texture and the thickness o f the endometrium as assessed by transvaginal sonography were prospectively evaluated in 123 patients undergoing I V F treatment. Three different types o f endometrial patterns could be distinguished: (A) an entirely homogeneous, hyperechogenic endometrium; (B) an intermediate type characterized by the same reflectivity o f ultrasound as the myornetrium, with a nonprominent or absent central echogenic line; and (C) a multilayered endometrium consisting o f prominent outer and midline hyperechogenic lines and inner hypoechogenic regions. On the day before oocyte retrieval, endometrial thickness was significantly greater in the group o f patients who achieved pregnancy than in the group who did not (8.7 +--0.4 vs 7.5 +- 0.2 ram, respectively; P < 0.01) and significantly more patients had multilayered, pattern C, endometrium (75% in pregnant women vs 42.4% in nonpregnant women; P < 0.01). N o pregnancy occurred when the endometrial thickness was less than 6 mm. When type C endometrium >-6 mm thick was seen, the pregnancy rate per embryo transfer was 39%. When type A or B endometrial pattern was seen, the negative predictive value f o r the occurrence o f pregnancy was 90.5%. Our results suggest that transvaginal sonographic evaluation o f endometrial texture and thickness may be an indicator o f the likelihood o f achieving pregnancy.
The introduction of the vaginal ultrasound transducer, with its high-frequency, high-resolution capability, represents a major advance in gynecologic sonography. Transvaginal ultrasound has focused mainly on measurement of follicular diameter for cycle monitoring, on its use as an improved modality for oocyte retrieval for in vitro fertilization (IVF), and on the early detection of intrauterine pregnancy by visualization of the gestational sac. Recently, we (1) and others (2) have shown that endometrial thickness and the amount of endometrial growth during controlled ovarian stimulation for IVF are significantly greater in conception cycles compared to nonconception cycles, a finding which may have potential predictive value with regard to implantation. In the present study, we attempt to extend these findings by examining both the thickness and the sonographic pattern of the endometrium in patients undergoing IVF. Different patterns of endometrial response in natural and stimulated cycles have been described (3-5) but a relationship between endometrial thickness and reflectivity to ultrasound has not been determined. Recently, Welker and coworkers (6) demonstrated that specific patterns of endometrial development can be distinguished by transvaginal ultrasound during IVF cycles and may be predictive of implantation. The objective of the present study was to determine, using vaginal sonography, whether a particular combination of endometrial pattern and growth during ovarian hyperstimulation for IVF could predict the subsequent occurrence of pregnancy.
KEY WORDS: transvaginal sonography; endometrial texture; endometrial thickness; in vitro fertilization.
i Division of Reproductive Science, Department of Obstetrics and Gynecology, The University of Toronto, Toronto, Canada. 2 Present address: Department of Obstetrics and Gynecology, Carmel Hospital, Haifa, Israel. 3 To whom correspondence should be addressed at 6-240 EN, Toronto General Hospital, 200 Elizabeth Street, Toronto, Ontario, Canada M5G 2C4.
0740-7769/90/0600-0146506.00/0 9 1990 Plenum Publishing Corporation
146
ENDOMETRIAL PATTERN AND THICKNESS DURING IVF
MATERIALS AND METHODS Subjects Endometrial thickness and pattern were prospectively studied during induction of ovulation in 123 consecutive patients undergoing treatment in the University of Toronto IVF program. The mean (-+SE) age of the women was 32.9 -+ 0.31 years, with a range from 25 to 39 years. Most of the patients (69.9%) had tubal infertility, 9.8% had endometriosis, and 8.9% were suffering from longstanding infertility of unknown etiology. Male factor was involved as the cause of infertility either as the sole problem or combined with tubal factor or endometriosis in 11.4% of patients. All patients included in the study underwent standardized ovarian follicular stimulation with clomiphene citrate (Serophene, Serono, Randolph, MA), 100 mg/day, cycle days 5-9, and human menopausal gonadotropins (hMG; Pergonal, Serono), 75 to 150 IU im starting on cycle day 5 or 6. Human chorionic gonadotropin (hCG; Profasi, Serono), 5000 IU im, was routinely used to trigger the final stage of follicular maturation. Oocyte retrieval and fertilization were performed using standard techniques in each patient without regard to the endometrial findings. Embryo transfer occurred 48 hr after retrieval in all patients. No luteal-phase support was used. Cycle Monitoring The response to treatment was monitored daily starting on day 9 of the menstrual cycle by daily determination of serum levels of estradiol (E2) and luteinizing hormone (LH) by radioimmunoassay (RIA) as previously described (7,8). Transvaginal ultrasound measurements of follicular diameter and endometrial thickness were recorded daily, starting on cycle day 10. Endometrial thickness was measured on the "frozen" ultrasound image obtained with a 7-MHz vaginal transducer (Bruel & Kjaer, Type 8538, Naerum, Denmark). The maximum thickness of endometrium on both sides of the midline was measured in the plane through the central longitudinal axis of the uterine body (Fig. I). Endometrial Pattern and Texture
Pictures of the endometrium were taken on the day after administration of hCG (the day prior to
147
ovum retrieval) using a Mitsubishi video copy processor (Type P-60, Japan) and were stored for later evaluation of endometrial pattern. Near and far gains were constant for all measurements. Both investigators independently assessed endometrial pattern in all patients with subjective agreement. Endometrial pattern was determined by comparing the reflectivity or gray-scale appearance of the endometrium to that of the adjacent myometrium. The changes in reflectivity varied from black or darker than the myometrium (hypoechogenic), to equal to the appearance of the myometrium (isoechogenic), to brighter than the myometrium (hyperechogenic). Three endometrial ultrasound patterns were distinguished. (A) An entirely homogeneous, hyperechogenic endometrium was characterized by increased reflectivity compared to the myometrium and consequently appeared brighter in its gray-scale appearance. The central echogenic line was not visualized (Fig. la). (B) An intermediate, isoechogenic pattern was characterized by the same reflectivity compared to the surrounding myometrium, with a nonprominent or absent central echogenic line (Fig. lb). (C) A multilayered "triple-line" endometrium consisted of prominent outer and central hyperechogenic lines and inner hypoechogenic or black regions (Fig. lc). Statistical Evaluation
The data obtained were compared by Student's t test for unpaired results and the chi-square test. P values 1.5 cm 6.2 - 0.3 5.9 -+ 0.3 No. oocytes per retrieval 5.4 --+ 0.4 5.5 --- 0.5 No. embryos per transfer 3.0 -+ 0.3 2.6 --- 0.4 a
Not significant.
NS a NS NS NS NS NS NS
Thirty-one of the 99 nonpregnant patients did not have an embryo transfer due to lack of fertilization, and in 3 others, no eggs were retrieved. In this group, 47% (16/34) displayed a pattern A endometrium. Significantly more patients with patterns A and B endometrium failed to achieve embryo transfer (23/63; 36.5%) than patients with pattern C (11/60; 18.3%; P < 0.05) (Table I). There was no difference in endometrial thickness between the group of patients with lack of fertilization and those who did not conceive after embryo transfer (7.3 -+ 0.4 and 7.7 - 0.2 mm, respectively).
DISCUSSION We found, using transvaginal sonography, that the presence of pattern A or B endometrium on the day after hCG administration (day prior to oocyte recovery) was associated with a negative predictive value for pregnancy of 90.5%. In addition, in over 60 IVF pregnancies obtained by us in the past year, none occurred in any patient found to have an endometrial thickness below 6 mm on the day after hCG. In contrast, the presence of pattern C endometrium of 6-mm thickness or greater on the day after hCG administration was associated with more than a 39% clinical pregnancy rate per embryo transfer in our IVF program. These data confirm the results of Welker et al. (6), who also reported higher pregnancy rates in women with endometrium displaying a triple-line pattern (pattern C) on the day of oocyte retrieval than in those with a hyperechogenic endometrium (pattern A). With these findings to date, we are quite confident that we can determine those patients who have a very poor chance of conceiving in a specific IVF cycle prior to the oocyte recovery. Furthermore, our results demonstrate that the endometrial pattern and thickness on the day before oocyte recovery, if unfavorable, are predictive of a negative outcome. The quality and state of preparation of the endometrium are two of several factors which may influence the success or failure of in vitro fertilization and embryo transfer (9). By using ultrasound examination, it is now possible to determine routinely and accurately, the thickness and texture of the endometrium (9-11). The transvaginal, highfrequency transducer enables even more detailed visualization of small changes in endometrial pattern and growth during the cycle. The criteria we studied, namely, endometrial pattern and thickness,
Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
150
GONEN AND CASPER
100 -
J
ill PREGNANT [ ] NONPREGNANT
I
80,
I'-
50'
zi l l r.3 iz: i,u n
40'
20'
.2-.3
.4-.5
.6-.7
.8-.9
1.0-1.1 1.2-1.3 1.4-1.5
ENDOMETRIAL THICKNESS (CM) Fig. 2. The occurrence of clinical pregnancy related to endometrial thickness. Percentage, on the Y axis, refers to the percentage of each group [pregnant cycles (hatched bars) and nonpregnant cycles (open bars)] which occurred at each level of endometrial thickness.
appear to be useful in prediction of pregnancy occurrence, especially when an unfavorable endometrial picture is visualized. This unfavorable pattern consists of three factors: (i) a relatively hyperechogenic or isoechogenic texture compared to the myometrium, (ii) the absence of hyperechogenic lines separating the endometrium from the myometrium and the absence or faint presence of a central echogenic line, and (iii) a thickness of less than 6 mm. In contrast, a favorable endometrial pattern for the subsequent occurrence of pregnancy appears as multilayered endometrium consisting of prominent outer and midline hyperechogenic lines and inner hypoechogenic or black regions and an endometrial thickness of at least 6 mm. This favorable pattern has been described by Forrest et al. (12) as typical of the proliferative endometrium in natural cycles and characterized by these authors as the appearance of "triple lines" running longitudinally in the center of the uterus. The central echogenic line is produced by the endometrial canal but the histologic structures that give rise to the echogenic outer lines are not known. Forrest et al. (12) believe that these outer lines may represent the basal layer of the endometrium or the interface between the endometrium and the myometrium. They
also describe clearly the relatively hypoechogenic regions between the two outer lines and the central line, which probably represents the functional layer of the endometrium. As endometrial development progresses during the late proliferative phase, the functional layer thickens and the distance between the outer two lines and the central echogenic line increases. Although, in general, a favorable, type C pattern of endometrial reflectivity is associated with a prominent endometrial thickness, we have occasionally found that an adequate increase in the thickness of the endometrium may fail to occur with this pattern. In all cases where the endometrial thickness was found to be less than 6 mm on the day prior to oocyte recovery, regardless of endometrial pattern, clinical pregnancy subsequently failed to occur in our program. We have previously demonstrated that fertilization rate and endometrial thickness are not correlated, and that these two measurements appear to be independent of each other (1). In contrast, Smith et al. (5) showed, using retrospective analysis, that an endometrial thickness of less than 5 mm (with transabdominal sonography) was associated with either failure to recover oocytes or the recovery of
Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
151
ENDOMETRIAL PATTERN AND THICKNESS DURING IVF
TYPE "A"
20-
15 !i!i~!i!~!i!!!~!~i~i!i!i
i~i!~!~:i!i!i!~!i!i~iil
10 ::::::::.::::,
:::::::::,:::
......................
~:~:~:~:~:~:~:~:~:~:~:~:7 ,::::,::::.::: .................. ::::::::::::::::::::::
::::::::::::::::::::::: ........................ ::::.::::::::: ::::::::::::
::::::::::::
.....................
~:~:~.::~:~::::
::::::::::::::::::::::
::::::::::::::::::::::::
0.2-0.3
0.4-0.5
:::::::::::::.::::::
0.6-0.7
0.8-019
1.o-1.1
1,2-1.3
1.4-1.5
TYPE "B" 10
U~ IJ.I
f
8
:::+:::,::::.:
.-I
r >.. r
:~::~::::.:,::,:.:,
6
4
Ii
O :it
2 ::,::::.::::.: ::::::::::::::::::::::::
0 20"
0.4-0.5
0.6-0.7
iiil
0.8-0.9
1.0-1.1
1.2-1.3
1.4-1.5
TYPE "C" 15 84
1 0 84
5'
0.4-0.5
0,6-0.7
0.8-0.9
1.0-1.1
ENDOMETRIAL THICKNESS
1.2-1.3
(cm)
Fig. 3. The relationship among endometrial pattern, endometrial thickness, and the occurrence of clinical pregnancy. (a) Pregnancy cycles (hatched bars) and nonpregnancy cycles (open bars) demonstrating endometrial pattern A (hyperechogenic) on the day after hCG administration related to endometrial thickness on the X axis. No pregnancy occurred with endometrial pattern A and a thickness less than 10 mm. (b) Pregnancy and nonpregnancy cycles in patients with endometrial pattern B (intermediate) on the day after hCG administration related to endometrial thickness. No pregnancy occurred with endometrial pattern B and thickness less than 8 mm. (c) Pregnancy and nonpregnancy cycles in patients with endometrial pattern C (multilayered) on the day after hCG. No pregnancy occurred if endometrial thickness was less than 6 mm.
immature oocytes with a lower fertilization rate. In the present study, the mean endometrial thickness in patients with lack of fertilization did not differ from the mean endometrial thickness of patients
who achieved fertilization and did not conceive, consistent with the findings of our previous study (1). In contrast, poor endometrial patterns (A and B) were associated with twice the incidence of fail-
Journal o f in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
152
GONEN AND CASPER
ure to fertilize than pattern C (multilayered pattern). These data suggest that changes in endometrial pattern during IVF stimulation may also be useful as predictive parameters for fertilization. In keeping with the findings of our present study, Welker et al. (6), also using transvaginal ultrasound, reported that the pattern of the endometrium is a predictive parameter for conception and suggested cancellation of oocyte collection or embryo replacement when a homogenous, hyperechogenic endometrium is present. In contrast to our results, these investigators did not find a correlation between endometrial thickness and occurrence of pregnancy. Glissant et al. (2) reported that endometrial thickness was significantly greater in conception compared to nonconception cycles but concluded that it was not possible to predict the probability of pregnancy with the use of endometrial thickness. We agree that the positive predictive value of endometrial thickness is low. However, our experience indicates that utilization of the combination of endometrial pattern and endometrial thickness has great negative predictive value for the subsequent occurrence of pregnancy. We believe that the findings of the present study will be helpful in decision-making regarding oocyte recovery, embryo replacement, and cryopreservation. If pattern A or B endometrium is seen or if the endometrial thickness is less than 6 mm with any pattern on the day after hCG administration, consideration should be given to oocyte recovery with transfer of frozen/thawed embryos during a subsequent cycle when more favorable endometrium is present. A few of the questions which remain to be answered include the following. (i) Does an unfavorable pattern or thickness of the endometrium recur in the same patient? (ii) Does the method of ovarian hyperstimulation contribute to the pattern and thickness of the endometrium? For example, could an antiestrogenic effect of clomiphene citrate causing estrogen receptor depletion in the endometrium be the cause of the unfavorable endometrial patterns and texture seen in the present study? (iii) Do unfavorable endometrial patterns or thicknesses occur in natural cycles? (iv) Can an unfavorable endometrial pattern or thickness be improved by hormonal manipulation or alteration of stimulation protocol? Further studies will be needed to address these and other questions concerning the preliminary findings of this report.
ACKNOWLEDGMENTS This work was supported by a research grant from the Medical Research Council of Canada and by a Fellowship Training Grant to Dr. Y. Gonen from Wyeth Pharmaceutical Company, Toronto, Ontario, Canada This paper was presented in part at the Postgraduate Course on Advanced IVF at the American Fertility Society Meeting, San Francisco, California, November 12, 1989. REFERENCES 1. Gonen Y, Casper RF, Jacobson W, Blankier J: Endometrial thickness and growth during ovarian stimulation: A possible predictor of implantation in in vitro fertilization. Fertil Steril 1989;52:446--450 2. Glissant A, de Mouzon J, Frydman R: Ultrasound study of the endometrium during in vitro fertilization cycles. Fertil Steril 1985;44:786-790 3. Fleischer AC, Kalemeris GC, Entman SS: Sonographic depiction of the endometrium during normal cycles. J Ultrasound Med Biol 1986;12:271-277 4. Fleischer AC, Pittway DE, Beard LA, Thieme GA, Bundy AL, James AE Jr, Wentz AC: Sonographic depiction of endometrial changes occurring with ovulation induction. J Ultrasound Med 1984;3:341-346 5. Smith B, Porter R, Ahuja K, Craft I: Ultrasonic assessment of endometrial changes in stimulated cycles in an in vitro fertilization and embryo transfer program. J Vitro Fert Embryo Transfer 1984;1:233-238 6. Welker BG, Gembruch U, Diedrich K, A1-Hasani S, Krebs D: Transvaginal sonography of the endometrium during ovum pickup in stimulated cycles for in vitro fertilization. J Ultrasound Med 1989;8:549--53 7. Gonen Y, Casper RF: Does transient hyperprolactinemia during ovarian hyperstimulation interfere with conception or pregnancy outcome. Fertil Steril 1989;51:1007-1010 8. Casper RF, Armstrong DT, Brown SE, Daniel SAJ, Yuzpe A, Erskine H: In vitro fertilization: Determination of follicular maturation for timing of human chorionic gonadotropin administration. Fertil Steril 1987;47:345-349 9. de Mouzon J, Lefevre B, Frydman R, Belaisch-Allert JC, Guillet-Rosso F, Testart J: Factors affecting human in vitro fertilization: a multifactorial study. Fertil Steril 1985;43:892896 10. Brant T, Levy E, Grant T, Marut E, Leland J: Endometrial echo and its significance in female infertility. Radiology 1985;157:225-229 Fleisher A, Herbert CM, Sacks GA, Wentz A, Entman S, James A: Sonography of the endometrium during conception and nonconception cycles of in vitro fertilization and embryo transfer. Fertil Steril 1986;46:442-447 12. Forrest TS, Elyaderni MK, Muilenburg MI, Bewtra C, Kable WT, Sullivan P: Cyclic endometrial changes: US assessment with histologic correlation. Radiology 1988;167: 233-237
Journal of in Vitro Fertilization and Embryo Transfer, Vol. 7, No. 3, 1990
