Human Reproduction vol.6 no.7 pp 959-962, 1991
Are there predictive criteria of complicated ovarian hyperstimulation in IVF?
A.Delvigne1, J.Vandromme, P.Barlow, B.Lejeune and F.Leroy 1VF Clinic and Human Reproduction Research Unit, Saint Pierre Hospital, Free University of Brussels, Belgium 'To whom correspondence should be addressed
Among 599 trials of in-vitro fertilization (TVF) treatment, complicated ovarian hyperstimulation (OHSS) was diagnosed in 14 cases (2.5%) on the basis of heavy abdominal discomfort and echographic findings (ascites, ovarian enlargement with cysts). Among eight hospitalized patients, four presented with a haemoconcentration and/or electrolytic disturbances. OHSS cases were compared with two control groups for a series of criteria: age, aetiology of infertility, total dose of human menopausal gonadotrophin (HMG), day of oocyte collection, oestradiol (E2) peak level, rate of E2 increase, number of oocytes, number of embryos transferred and embryonic vitality scores. Comparison with a random group of normal IVF trials showed a significant difference for the following parameters: E2 peak level and rate of increase, E^/dose of HMG, EVday of egg collection and number of oocytes. When OHSS cases were compared to another control group consisting only of high E2 responders (peak E2 > 2700 pg/ml), no significant difference was found for any of the above-mentioned criteria. In view of this lack of predictive power of individual criteria, stepwise discriminant analysis was applied, showing that this method might provide a predictive mathematical function for evaluating the risk of OHSS before human chorionic gonadotrophin (HCG) administration. Such a formula, however, should be validated by a multicentric study in which a greater number of OHSS cases would be tested.
are observed and the patient may present with vomiting and/or diarrhoea (grade 4). In addition to these symptoms, the severe form is characterized by accumulation of extravascular fluid (grade 5) (ascites, hydrothorax, pericardia] effusion) entailing hypovolaemia, haemoconcentration, oliguria and electrolytic disturbances (grade 6). In the most serious cases thromboembolic accidents may ensue. OHSS is thus an iatrogenic and sometimes life-threatening complication which may require intensive care and protracted hospitalization (Figueroa-Cases, 1958; Muller, 1962; Moses etal., 1965). In in-vitro fertilization (IVF) treatment, mild ovarian hyperstimulation has become a primary goal, allowing the recruitment of a cohort of mature follicles. However, in spite of its importance for infertile couples, IVF treatment is never a vital necessity nor can it be claimed to improve physical health. Therefore, care to prevent a major risk such as severe OHSS is especially important in this context. On the other hand, in view of the heavy emotional and financial investments in IVF trials, the rate of cycle cancellation due to a poor ovarian response should be kept as low as possible. Since it is difficult to foresee the amplitude of the individual ovarian response and adapt the dosage of gonadotrophin treatment accordingly, there is a need to define precise criteria which would enable clinicians to detect cases at risk of developing a complicated OHSS.
Materials and methods
Introduction
The present study deals with 599 IVF cycles performed under buserelin — human menopausal gonadotrophin —human chorionic gonadotrophin (HMG —HCG) treatment. Complicated ovarian hyperstimulation was usually suspected when a patient complained of heavy abdominal discomfort a few days after embryo replacement. Moderate and severe cases were defined using the criteria suggested by Rabau et al. (1967). According to this classification, 14 patients suffered from a complicated OHSS, among which eight were considered to be in need of hospitalization.
Medical induction of ovulation has been applied since the late thirties (Cole and Hart, 1930). A series of complications may arise from this treatment which are described under the genera] heading of 'ovarian hyperstimulation syndrome' (OHSS). This syndrome is conventionally classifed as mild, moderate or severe, each class being again subdivided into two grades of intensity (Rabau etal., 1967). The mild form (grades 1 and 2) corresponds to minor clinical signs such as discrete enlargement of the ovaries coexisting with small and sometimes palpable cysts. In moderate cases, abdominal distention and large cystic ovaries (grade 3)
Stimulation treatment The gonadotrophin releasing hormone (GnRH) agonist buserelin (Suprefact, Hoechst, Frankfurt) was administered by intranasal spray from the first day of menses at a dose of 100 /ig given four times per day. HMG (Humegon, Organon, Oss or Pergonal, Serono, Milano) was injected daily from the fifth day of the cycle. The daily dose of gonadotrophin treatment ranged from 75 IU to 450 IU and was defined according to previously reached oestradiol (E2) levels when patients had already undergone an
Key words: IVF/complicated ovarian (OHSS)/oestradiol/peak oestradiol
Oxford University Press.
hyperstimulation
959
A.Delvigne et al.
earlier FVF trial. Moreover, it was adjusted to current ovarian response in all cases. The daily dose of attack was initiated at 150IU per day when ovarian reactivity had not been previously documented. A dose of 10 000 IU HCG (Pregnyl, Organon) was administered when a diameter of 20-22 mm measured by echography for the largest follicles and an E2 level of 300 jig per ml per mature follicle were obtained. Other IVF procedures used by our team have been previously described in detail (Lejeune et al., 1986). Study groups The 14 OHSS cases were compared to two different control groups for a series of parameters. The first set of 20 controls was selected at random from the general population of IVF trials which did not entail clinical symptoms of OHSS. Since E^ levels > 2700 pg/ml had almost always been observed in clinical OHSS cases, the latter were also compared to 14 randomly selected IVF cycles showing similarly high E2 values but without having developed OHSS symptoms (total population of such cases: 158). OHSS cases included six non-hospitalized and eight hospitalized cases. Hospitalization was decided on the basis of the severity of abdominal pain and echographic findings (size of ovaries > 7 cm in diameter and presence of ascites). In this latter group.
patients were classified as follows: grade 6, four cases; grade 5, three cases; grade 4, one case. Among non-hospitalized patients the distribution was: grade 3, four cases; grade 5, two cases. Statistical analysis The data were analysed using an SPSS-PC Package on a Compaq 386 Computer. The first control group was selected at random among the whole IVF population by applying an SPSS-random sample program, based on a random number generator. This group was secondarily compared with the whole IVF population. The validity of this sample was confirmed by the absence of any statistical difference from the whole population for the following criteria: age, number of HMG ampoules, oestradiol peak level, oestradiol/dose of HMG, day of oocyte retrieval, E2/day of oocyte retrieval, number of eggs retrieved, E2/eggs retrieved, number of embryos transferred, vitality score of all embryos and of transferred embryos, aetiology of infertility and pregnancy rate. The second control group was established by pairing each OHSS case with a cycle from the remaining IVF population showing a similar high E2 level. Statistical comparison was effected by the Mann-Witney non-
Table I. Comparison between OHSS cases and normal controls (mean values, range given in parentheses)
Number of cases Age (years) Number HMG ampoules (75 lU/amp.) E2 peak level (pg/ml) E2 rate of increase1 Ej/dose of HMG Day of oocyte retrieval (OPU) E2/day of OPU Number of eggs retrieved E2/egg retrieved Number of embryos transferred Vitality score of all embryosb Vitality score of transferred embryos
Controls
OHSS cases
r-value
20 32 4 (25-39) 30.3 (12-60) 1979 (978-5777) 0.3O (0.17-0.41) 81 (19-208) 162 (12-21) 124 (57-361) 7.4 (1-19) 315 (140-918) 2.9 (0-3) 11 4 (0-28) 14.9 (0-18)
14 30.5 (22-39) 22.6 (10-46) 4253 (2741-6038) 0.39 (0.24-0.55) 223 (98-503) 167 (15-19) 257 (144-358) 14 1 (8-30) 297 (167-572) 2.4 (0-3) 10 7 (0-45) 16 4 (0-18)
NS NS < < < NS < < NS NS NS NS
X
10- 4 10" 2 10" 3
X X
10-' 10- 3
X X
"Ej growth rate is defined as the slope of the regression line of the semi-logarithmic plot of E2 levels during the last 4 days of ovarian stimulation (Dimfeld et al., 1985). ""Vitality scores are defined according to the amount of anucleate fragments and speed of development (Puissant et al , 1987). NS. not significant.
Table II. Comparison between OHSS cases and patients with similar Ei peak levels but without OHSS symptoms (mean values, range in parentheses)
Number of cases Age (years) Number of HMG ampoules E2 rate of increase1 E2/dose of HMG Day of oocyte retrieval (OPU) E^/day of oocyte OPU Number of eggs retrieved E2/egg retrieved Number of embryos transferred Vitality score of all embryosb Vitality score of transferred embryos
OHSS
E2 paired group
P-value
14 30.5 (22-29) 22.6 (10-46) 0.39 (0.24-0 55) 223 (89-503) 16.7 (15-19) 257 (144-358) 14.1 (8-30) 297 (167-572) 2.4 (0-3) 10.7 (0-45) 16.4 (0-18)
14 30.8 (25-39) 26 0 (14-40) 0.37 (0.25-0 54) 184 (95-358) 159 (14-19) 272 (164-399) 15.0 (2-32) 383 (127-1371) 2.7 (0-4) 10.4 (0-42) 16 9 (0-18)
NS NS NS NS NS NS NS NS NS NS NS
•Ej rate of increase is defined as the slope of the regression line of the semi-logarithmic plot of E-, levels during the last 4 days of ovarian stimulation (Dirnfeld et al.. 1985) Vitality scores are defined according to the presence of anucleate fragments and speed of development (Puissant et al.. 1987). NS, not significant
960
Evaluation of predictive criteria of OHSS in IVF
parametric test or by Pearson's chi-square test for pregnancy rates. Results The comparison between OHSS cases and normal controls is given in Table I. As expected, the peak E2 levels and number of oocytes recovered were significantly higher in the OHSS group. Although oocyte retrieval did not occur on different days of the cycle, the peak level of E2 as related to the day of oocyte retrieval was found to be significandy lower in controls. The E^ rate of increase, that is, the slope of the regression line of the semi-logarithmic plot of Ej levels during the last 4 days of stimulation (Dirnfield et al., 1985) was significandy higher in OHSS cases. The total dose of HMG tended to be greater in the control group but die difference did not reach statistical significance. However, the ratio between the peak E2 level and the number of HMG ampoules was significantly higher in OHSS cases. The patient's age and Ej secretion per oocyte retrieved did not significantly differ between groups. The total, clinical and ongoing pregnancy rates were not significantly different. The numbers of embryos transferred and their vitality scores (Puissant et al., 1987) also did not differ significantly. No statistical difference (8/20 versus 6/14) was observed between the two groups in terms of total pregnancies. However, pregnancy rates appeared to be greater in OHSS cases (6/14) than in the overall IVF population (207/758). When studying only the group of hospitalized OHSS cases, the same differences were found. In addition, the number of HMG ampoules appeared to be significantly lower in OHSS cases than in controls P < 0.04). When comparing OHSS cases to the selected control population which had also shown high E-> peak levels, no significant difference remained with respect to the above mentioned criteria (Table H). A high proportion of cases with 'minor endocrinopathy' was observed in the OHSS group though this difference was not significant. These cases were defined as showing abnormally high levels of androgens (i.e. testosterone > 55 ng/ml, androstenedione > 2 ng/ml, dihydroepiandrosterone sulphate > 3000 ng/ml) or of luteinizing hormone (LH); (values > 15 mUI/ml outside the ovulatory peak) or high basal prolactin (PRL) values (PRL > 500 /iU/ml).
Fig. 1. Distribution of oestradiol peak levels (pg/ml) in controls (D) and OHSS cases ( • ) . n = 599.
The group of OHSS cases that we studied presented a relative paucity of cases of typical multicystic ovarian disease (PCO). In the OHSS group, this pathology was presented by only two cases of pure endocrinopathy and by one of mixed aetiology (male plus endocrine factors). PCO was defined as an endocrine profile showing a luteinizing hormone/follicle stimulating hormone (LH/FSH) ratio above two (Givens et al., 1976), elevated androstenedione and testosterone levels (Bardin and Lipsett, 1967; Devane et al., 1975) associated widi characteristic echographic findings (Orsini et ai, 1985).
Discussion Definition of the ovarian hyperstimulation syndrome raises conceptual problems. In our work, OHSS cases were primarily defined as patients who, a few days after embryo transfer, complained of heavy abdominal pain. Because of die subjectiveness of this symptom, other parameters were used, such as clinical evidence of abdominal tension and ascites as well as echographic findings (size of ovaries and ascites). The latter parameters were used to decide whether patients had to be hospitalized or put under ambulatory surveillance. Among eight patients who were hospitalized, four grade 6 cases presented haemoconcentration (haematocrit > 45%) and/or electrolytic disturbances (K+ either < 3.5 meq/1 or > 4.5 meq/1)- The validity of our hospitalization criteria seemed to be confirmed by the observation that no such biological anomalies were detected among ambulatory cases. Since the evaluation of pain is highly subjective, it may be wondered how many patients less sensitive to pain escaped detection and to what extent the occurrence of abdominal discomfort, increase of ovarian size and accumulation of some fluid in the Douglas pouch are acceptable after an IVF trial. In our recent experience, > 50% of FVF patients exhibit one or other of these features. It appears mat in the context of IVF and in view of the availability of vaginal echographic probes, which can detect as little as a few millilitres of fluid, classical criteria for OHSS should be redefined. This would involve systematic echographic screening after embryo transfer of every IVF patient in order to try and detect cases at risk of severe OHSS. As already described (Schenker and Weinstein, 1978; Haning et al., 1983; Teietal., 1985), patients who fulfilled OHSS criteria showed significantly elevated oestradiol levels and also higher numbers of oocytes than other IVF cases. Also, the study of the rate of E2 increase and of the relationship between the E2 peak and day of oocyte retrieval indicates that, on average, high E2 levels are reached more rapidly in these patients. The higher speed of die oestrogenic response to lower doses of HMG in OHSS cases confirms that the ovaries of these patients are hypersensitive to exogenous gonadotrophins (Golan et al., 1988). It has been suggested that the risk of OHSS is greater when a large number of small ( < 9 mm) and medium-sized (9—15 mm) follicles, versus only a few mature ones, are present on the day of HCG administration (Blankstein et al., 1987). If so, die ratio of E3 to the number of eggs retrieved, should be higher in OHSS-threatened cases, since small follicles, while contributing to high E2 levels, are difficult to puncture effi961
A.Delvigne et al.
ciently. In our patients, however, this ratio was not higher among OHSS cases. Although the E2 peak level and rate of increase, ratios of E2 versus dose of HMG, E2 versus day of retrieval and number of oocytes obtained are statistically different in OHSS and control groups, none of these parameters is discriminant by itself because of the large distribution of values in normal cases. The example of E2 peak levels clearly illustrates this difficulty (Figure 1), since the right-hand tail of the control distribution entirely overlaps the OHSS distribution. Therefore, it is difficult to make the decision of abandoning an IVF cycle at the risk of entailing severe OHSS, only on the basis of high Ej levels. A predictive approach is that of Tibi et al. (1989), which is based on pelvic ultrasound examination before ovarian stimulation for IVF. These authors reported that the presence of 10 or more small follicles from 4 to 8 mm in diameter in one or both ovaries was associated with a 75 % rate of polycystic ovarian response and a 42.8% rate of biological hyperstimulation (sensitivity 50%, specificity, 91.4%). An additional approach would be to evaluate the OHSS risk through discriminant analysis, enabling the derivation of a predictive function based on several criteria studied during ovarian stimulation, as in the present report. Although the number of OHSS cases that we observed is too small to derive firm conclusions, discriminant evaluation of the data was performed as a preliminary study. The following discriminant score was obtained: A = - 2 . 9 8 + (0.66 x 103 E J + (0.95 x 10"' x number of eggs retrieved). Taking A > 0.2, the rates of false negative and false positive cases are 7% and 10%, respectively, with an overall predictability of 91%. Applying the same approach to hospitalized patients only, reduces the false negative and false positive cases to 5% and 0% respectively. However, the group that we have studied can at most be used as a 'learning set' for deriving a hypothetical predictive function. The latter should be applied prospectively to a large population of IVF trials before deriving firm conclusions (Romeder, 1973). In view of the relatively small number of cases occurring in each centre, this type of study should be carried out on a multicentric basis. References Bardin,C.W. and Lipsett.M.B. (1967) Testosterone and androstenedione blood production rates in normal women and women with idiopathic hirsutism or polycystic ovaries. J. Clin. Invest., 46, 891—902. Blankstein,J., Shalev.J., Saadon.T., Kukia,E.E., Rabinovici.J.. Pariente,C, Lunenfeld,B., Serr,D.M. and Mashiach.S. (1987) Ovarian hyperstimulation syndrome: prediction by number and size of preovulating ovarian follicles. Fertil. Steril., 47, 597 — 602. Cole.H.H. and Hart.G.H. (1930) The potency of blood serum of mare in progressive stage of pregnancy in effecting the sexual maturity of the immature rat. Am. J. Physiol., 93, 57. Devane,G.W., Czekala,M., Judd,H.L. and Yen,S.C. (1975) Circulating gonadotropins, estrogens and androgens in polycystic ovarian disease Am. J. Obstet. GynecoL, 121, 496-500. Dirnfeld.M., Lejeune.B., Camus.M., Vekemans,M. and Leroy.F. (1985) Growth rate of follicular estrogen secretion in relation to the outcome of in vitro fertilization and embryo replacement. Fertil. Steril., 43, 379-384. Figueroa-Cases,P. (1958) Reaccion ovarica monstruosa a las
962
gonadotropines a proposito de un caso fatal. Ann. drug., 23, 116-118. Givens,J.R., Andersen,R.N., Umstot.E.S. and Wiser,W.L. (1976) Clinical findings and hormonal responses in patients with polycystic ovarian disease with normal versus elevated LH levels. Obstet. GynecoL, 47, 338-394. Golan,A., Ron-E1,R., Herman,A., Weimraub,Z., Soffer.Y. and Caspi.E. (1988) Ovarian hyperstimulation syndrome following r>Trp-6 luteinizing hormone-releasing hormone microcapsules and metropin for in vitro fertilization. Fertil. Steril., 50, 912-916. Haning.R.V., Austin,C.W., Carlson,I.H., Kuzma,D.L., Shapiro.S.S. and Zweibel.W.J. (1983) Plasma estradiol is superior to ultrasound and urinary estriol glycuronide as a predictor of ovarian hyperstimulation during induction of ovulation with menotropins. Fertil. Steril., 40, 31-36. Lejeune,B., Degueldre.M., Camus,M., Vekemans,M., Opsomer,L. and Leroy,F. (1986) In vitro fertilization and embryo transfer as related to endogenous luteinizing hormone rise or human chorionic gonadotropin. Fertil. Steril., 45, 377-383. Moses,M., Bogowsky,H., Anteby.E., Lunenfeld.B., Rabau.E., Serr,D., David,A. and Salomy,M. (1965) Thromboembolic phenomena after ovarian stimulation with human menopausal gonadotrophins. Lancet, ii, 1213-1215. Muller,P. (1962) In Beclere.C. (ed.), Gonadotrophines en Gvnecologie. Masson, Paris, p. 137. Orsini,L.F., Venturoh.S., Lurosso.R., Pluchinotta.V., Paradisi,R. and Bovicelli,L. (1985) Ultrasonic findings in polycystic ovarian disease.
Fertil. Steril., 43, 709-714. Puissant,F., Van Rysselberge,M., Barlow,P., Deweze,J. and Leroy.F. (1987) Embryo scoring as a prognostic tool in IVF treatment. Hum. Reprod., 8, 705-708. Rabau,E., David,A., Serr.D.M., Mashiach,S. and Lunenfeld,B. (1967) Human menopausa] gonadotropins for anovulation and sterility. Am. J. Obstet. GynecoL, 98, 92. Romeder J.M. (1973) Mtthode et Programmes d'Analyse Discriminante. Dunod, Paris, p. 66. SchenkerJ.G. and Weinstein.D. (1978) Ovarian hyperstimulation syndrome: a current survey. Fertil. Steril., 30, 255—268. Tal,J., Paz,B., Samberg.I., Lazaroun.N. and Sharf.M. (1985) Ultrasonographic and clinical correlates of metropin versus sequential clomiphene citrate: metropin for induction of ovulation. Fertil. Steril., 44, 342. Tibi,C, Alvarez.S., Comet,D., Antoine,J.M., Gomez,A.C. and SalatBaroux,J. (1989) Prediction des hyperstimulations ovariennes. Contracept. Fertil. Sexual.. 17, 751-752 Received on January 14, 1991; accepted on April 25, 1991
Are there predictive criteria of complicated ovarian hyperstimulation in IVF?
A.Delvigne1, J.Vandromme, P.Barlow, B.Lejeune and F.Leroy 1VF Clinic and Human Reproduction Research Unit, Saint Pierre Hospital, Free University of Brussels, Belgium 'To whom correspondence should be addressed
Among 599 trials of in-vitro fertilization (TVF) treatment, complicated ovarian hyperstimulation (OHSS) was diagnosed in 14 cases (2.5%) on the basis of heavy abdominal discomfort and echographic findings (ascites, ovarian enlargement with cysts). Among eight hospitalized patients, four presented with a haemoconcentration and/or electrolytic disturbances. OHSS cases were compared with two control groups for a series of criteria: age, aetiology of infertility, total dose of human menopausal gonadotrophin (HMG), day of oocyte collection, oestradiol (E2) peak level, rate of E2 increase, number of oocytes, number of embryos transferred and embryonic vitality scores. Comparison with a random group of normal IVF trials showed a significant difference for the following parameters: E2 peak level and rate of increase, E^/dose of HMG, EVday of egg collection and number of oocytes. When OHSS cases were compared to another control group consisting only of high E2 responders (peak E2 > 2700 pg/ml), no significant difference was found for any of the above-mentioned criteria. In view of this lack of predictive power of individual criteria, stepwise discriminant analysis was applied, showing that this method might provide a predictive mathematical function for evaluating the risk of OHSS before human chorionic gonadotrophin (HCG) administration. Such a formula, however, should be validated by a multicentric study in which a greater number of OHSS cases would be tested.
are observed and the patient may present with vomiting and/or diarrhoea (grade 4). In addition to these symptoms, the severe form is characterized by accumulation of extravascular fluid (grade 5) (ascites, hydrothorax, pericardia] effusion) entailing hypovolaemia, haemoconcentration, oliguria and electrolytic disturbances (grade 6). In the most serious cases thromboembolic accidents may ensue. OHSS is thus an iatrogenic and sometimes life-threatening complication which may require intensive care and protracted hospitalization (Figueroa-Cases, 1958; Muller, 1962; Moses etal., 1965). In in-vitro fertilization (IVF) treatment, mild ovarian hyperstimulation has become a primary goal, allowing the recruitment of a cohort of mature follicles. However, in spite of its importance for infertile couples, IVF treatment is never a vital necessity nor can it be claimed to improve physical health. Therefore, care to prevent a major risk such as severe OHSS is especially important in this context. On the other hand, in view of the heavy emotional and financial investments in IVF trials, the rate of cycle cancellation due to a poor ovarian response should be kept as low as possible. Since it is difficult to foresee the amplitude of the individual ovarian response and adapt the dosage of gonadotrophin treatment accordingly, there is a need to define precise criteria which would enable clinicians to detect cases at risk of developing a complicated OHSS.
Materials and methods
Introduction
The present study deals with 599 IVF cycles performed under buserelin — human menopausal gonadotrophin —human chorionic gonadotrophin (HMG —HCG) treatment. Complicated ovarian hyperstimulation was usually suspected when a patient complained of heavy abdominal discomfort a few days after embryo replacement. Moderate and severe cases were defined using the criteria suggested by Rabau et al. (1967). According to this classification, 14 patients suffered from a complicated OHSS, among which eight were considered to be in need of hospitalization.
Medical induction of ovulation has been applied since the late thirties (Cole and Hart, 1930). A series of complications may arise from this treatment which are described under the genera] heading of 'ovarian hyperstimulation syndrome' (OHSS). This syndrome is conventionally classifed as mild, moderate or severe, each class being again subdivided into two grades of intensity (Rabau etal., 1967). The mild form (grades 1 and 2) corresponds to minor clinical signs such as discrete enlargement of the ovaries coexisting with small and sometimes palpable cysts. In moderate cases, abdominal distention and large cystic ovaries (grade 3)
Stimulation treatment The gonadotrophin releasing hormone (GnRH) agonist buserelin (Suprefact, Hoechst, Frankfurt) was administered by intranasal spray from the first day of menses at a dose of 100 /ig given four times per day. HMG (Humegon, Organon, Oss or Pergonal, Serono, Milano) was injected daily from the fifth day of the cycle. The daily dose of gonadotrophin treatment ranged from 75 IU to 450 IU and was defined according to previously reached oestradiol (E2) levels when patients had already undergone an
Key words: IVF/complicated ovarian (OHSS)/oestradiol/peak oestradiol
Oxford University Press.
hyperstimulation
959
A.Delvigne et al.
earlier FVF trial. Moreover, it was adjusted to current ovarian response in all cases. The daily dose of attack was initiated at 150IU per day when ovarian reactivity had not been previously documented. A dose of 10 000 IU HCG (Pregnyl, Organon) was administered when a diameter of 20-22 mm measured by echography for the largest follicles and an E2 level of 300 jig per ml per mature follicle were obtained. Other IVF procedures used by our team have been previously described in detail (Lejeune et al., 1986). Study groups The 14 OHSS cases were compared to two different control groups for a series of parameters. The first set of 20 controls was selected at random from the general population of IVF trials which did not entail clinical symptoms of OHSS. Since E^ levels > 2700 pg/ml had almost always been observed in clinical OHSS cases, the latter were also compared to 14 randomly selected IVF cycles showing similarly high E2 values but without having developed OHSS symptoms (total population of such cases: 158). OHSS cases included six non-hospitalized and eight hospitalized cases. Hospitalization was decided on the basis of the severity of abdominal pain and echographic findings (size of ovaries > 7 cm in diameter and presence of ascites). In this latter group.
patients were classified as follows: grade 6, four cases; grade 5, three cases; grade 4, one case. Among non-hospitalized patients the distribution was: grade 3, four cases; grade 5, two cases. Statistical analysis The data were analysed using an SPSS-PC Package on a Compaq 386 Computer. The first control group was selected at random among the whole IVF population by applying an SPSS-random sample program, based on a random number generator. This group was secondarily compared with the whole IVF population. The validity of this sample was confirmed by the absence of any statistical difference from the whole population for the following criteria: age, number of HMG ampoules, oestradiol peak level, oestradiol/dose of HMG, day of oocyte retrieval, E2/day of oocyte retrieval, number of eggs retrieved, E2/eggs retrieved, number of embryos transferred, vitality score of all embryos and of transferred embryos, aetiology of infertility and pregnancy rate. The second control group was established by pairing each OHSS case with a cycle from the remaining IVF population showing a similar high E2 level. Statistical comparison was effected by the Mann-Witney non-
Table I. Comparison between OHSS cases and normal controls (mean values, range given in parentheses)
Number of cases Age (years) Number HMG ampoules (75 lU/amp.) E2 peak level (pg/ml) E2 rate of increase1 Ej/dose of HMG Day of oocyte retrieval (OPU) E2/day of OPU Number of eggs retrieved E2/egg retrieved Number of embryos transferred Vitality score of all embryosb Vitality score of transferred embryos
Controls
OHSS cases
r-value
20 32 4 (25-39) 30.3 (12-60) 1979 (978-5777) 0.3O (0.17-0.41) 81 (19-208) 162 (12-21) 124 (57-361) 7.4 (1-19) 315 (140-918) 2.9 (0-3) 11 4 (0-28) 14.9 (0-18)
14 30.5 (22-39) 22.6 (10-46) 4253 (2741-6038) 0.39 (0.24-0.55) 223 (98-503) 167 (15-19) 257 (144-358) 14 1 (8-30) 297 (167-572) 2.4 (0-3) 10 7 (0-45) 16 4 (0-18)
NS NS < < < NS < < NS NS NS NS
X
10- 4 10" 2 10" 3
X X
10-' 10- 3
X X
"Ej growth rate is defined as the slope of the regression line of the semi-logarithmic plot of E2 levels during the last 4 days of ovarian stimulation (Dimfeld et al., 1985). ""Vitality scores are defined according to the amount of anucleate fragments and speed of development (Puissant et al , 1987). NS. not significant.
Table II. Comparison between OHSS cases and patients with similar Ei peak levels but without OHSS symptoms (mean values, range in parentheses)
Number of cases Age (years) Number of HMG ampoules E2 rate of increase1 E2/dose of HMG Day of oocyte retrieval (OPU) E^/day of oocyte OPU Number of eggs retrieved E2/egg retrieved Number of embryos transferred Vitality score of all embryosb Vitality score of transferred embryos
OHSS
E2 paired group
P-value
14 30.5 (22-29) 22.6 (10-46) 0.39 (0.24-0 55) 223 (89-503) 16.7 (15-19) 257 (144-358) 14.1 (8-30) 297 (167-572) 2.4 (0-3) 10.7 (0-45) 16.4 (0-18)
14 30.8 (25-39) 26 0 (14-40) 0.37 (0.25-0 54) 184 (95-358) 159 (14-19) 272 (164-399) 15.0 (2-32) 383 (127-1371) 2.7 (0-4) 10.4 (0-42) 16 9 (0-18)
NS NS NS NS NS NS NS NS NS NS NS
•Ej rate of increase is defined as the slope of the regression line of the semi-logarithmic plot of E-, levels during the last 4 days of ovarian stimulation (Dirnfeld et al.. 1985) Vitality scores are defined according to the presence of anucleate fragments and speed of development (Puissant et al.. 1987). NS, not significant
960
Evaluation of predictive criteria of OHSS in IVF
parametric test or by Pearson's chi-square test for pregnancy rates. Results The comparison between OHSS cases and normal controls is given in Table I. As expected, the peak E2 levels and number of oocytes recovered were significantly higher in the OHSS group. Although oocyte retrieval did not occur on different days of the cycle, the peak level of E2 as related to the day of oocyte retrieval was found to be significandy lower in controls. The E^ rate of increase, that is, the slope of the regression line of the semi-logarithmic plot of Ej levels during the last 4 days of stimulation (Dirnfield et al., 1985) was significandy higher in OHSS cases. The total dose of HMG tended to be greater in the control group but die difference did not reach statistical significance. However, the ratio between the peak E2 level and the number of HMG ampoules was significantly higher in OHSS cases. The patient's age and Ej secretion per oocyte retrieved did not significantly differ between groups. The total, clinical and ongoing pregnancy rates were not significantly different. The numbers of embryos transferred and their vitality scores (Puissant et al., 1987) also did not differ significantly. No statistical difference (8/20 versus 6/14) was observed between the two groups in terms of total pregnancies. However, pregnancy rates appeared to be greater in OHSS cases (6/14) than in the overall IVF population (207/758). When studying only the group of hospitalized OHSS cases, the same differences were found. In addition, the number of HMG ampoules appeared to be significantly lower in OHSS cases than in controls P < 0.04). When comparing OHSS cases to the selected control population which had also shown high E-> peak levels, no significant difference remained with respect to the above mentioned criteria (Table H). A high proportion of cases with 'minor endocrinopathy' was observed in the OHSS group though this difference was not significant. These cases were defined as showing abnormally high levels of androgens (i.e. testosterone > 55 ng/ml, androstenedione > 2 ng/ml, dihydroepiandrosterone sulphate > 3000 ng/ml) or of luteinizing hormone (LH); (values > 15 mUI/ml outside the ovulatory peak) or high basal prolactin (PRL) values (PRL > 500 /iU/ml).
Fig. 1. Distribution of oestradiol peak levels (pg/ml) in controls (D) and OHSS cases ( • ) . n = 599.
The group of OHSS cases that we studied presented a relative paucity of cases of typical multicystic ovarian disease (PCO). In the OHSS group, this pathology was presented by only two cases of pure endocrinopathy and by one of mixed aetiology (male plus endocrine factors). PCO was defined as an endocrine profile showing a luteinizing hormone/follicle stimulating hormone (LH/FSH) ratio above two (Givens et al., 1976), elevated androstenedione and testosterone levels (Bardin and Lipsett, 1967; Devane et al., 1975) associated widi characteristic echographic findings (Orsini et ai, 1985).
Discussion Definition of the ovarian hyperstimulation syndrome raises conceptual problems. In our work, OHSS cases were primarily defined as patients who, a few days after embryo transfer, complained of heavy abdominal pain. Because of die subjectiveness of this symptom, other parameters were used, such as clinical evidence of abdominal tension and ascites as well as echographic findings (size of ovaries and ascites). The latter parameters were used to decide whether patients had to be hospitalized or put under ambulatory surveillance. Among eight patients who were hospitalized, four grade 6 cases presented haemoconcentration (haematocrit > 45%) and/or electrolytic disturbances (K+ either < 3.5 meq/1 or > 4.5 meq/1)- The validity of our hospitalization criteria seemed to be confirmed by the observation that no such biological anomalies were detected among ambulatory cases. Since the evaluation of pain is highly subjective, it may be wondered how many patients less sensitive to pain escaped detection and to what extent the occurrence of abdominal discomfort, increase of ovarian size and accumulation of some fluid in the Douglas pouch are acceptable after an IVF trial. In our recent experience, > 50% of FVF patients exhibit one or other of these features. It appears mat in the context of IVF and in view of the availability of vaginal echographic probes, which can detect as little as a few millilitres of fluid, classical criteria for OHSS should be redefined. This would involve systematic echographic screening after embryo transfer of every IVF patient in order to try and detect cases at risk of severe OHSS. As already described (Schenker and Weinstein, 1978; Haning et al., 1983; Teietal., 1985), patients who fulfilled OHSS criteria showed significantly elevated oestradiol levels and also higher numbers of oocytes than other IVF cases. Also, the study of the rate of E2 increase and of the relationship between the E2 peak and day of oocyte retrieval indicates that, on average, high E2 levels are reached more rapidly in these patients. The higher speed of die oestrogenic response to lower doses of HMG in OHSS cases confirms that the ovaries of these patients are hypersensitive to exogenous gonadotrophins (Golan et al., 1988). It has been suggested that the risk of OHSS is greater when a large number of small ( < 9 mm) and medium-sized (9—15 mm) follicles, versus only a few mature ones, are present on the day of HCG administration (Blankstein et al., 1987). If so, die ratio of E3 to the number of eggs retrieved, should be higher in OHSS-threatened cases, since small follicles, while contributing to high E2 levels, are difficult to puncture effi961
A.Delvigne et al.
ciently. In our patients, however, this ratio was not higher among OHSS cases. Although the E2 peak level and rate of increase, ratios of E2 versus dose of HMG, E2 versus day of retrieval and number of oocytes obtained are statistically different in OHSS and control groups, none of these parameters is discriminant by itself because of the large distribution of values in normal cases. The example of E2 peak levels clearly illustrates this difficulty (Figure 1), since the right-hand tail of the control distribution entirely overlaps the OHSS distribution. Therefore, it is difficult to make the decision of abandoning an IVF cycle at the risk of entailing severe OHSS, only on the basis of high Ej levels. A predictive approach is that of Tibi et al. (1989), which is based on pelvic ultrasound examination before ovarian stimulation for IVF. These authors reported that the presence of 10 or more small follicles from 4 to 8 mm in diameter in one or both ovaries was associated with a 75 % rate of polycystic ovarian response and a 42.8% rate of biological hyperstimulation (sensitivity 50%, specificity, 91.4%). An additional approach would be to evaluate the OHSS risk through discriminant analysis, enabling the derivation of a predictive function based on several criteria studied during ovarian stimulation, as in the present report. Although the number of OHSS cases that we observed is too small to derive firm conclusions, discriminant evaluation of the data was performed as a preliminary study. The following discriminant score was obtained: A = - 2 . 9 8 + (0.66 x 103 E J + (0.95 x 10"' x number of eggs retrieved). Taking A > 0.2, the rates of false negative and false positive cases are 7% and 10%, respectively, with an overall predictability of 91%. Applying the same approach to hospitalized patients only, reduces the false negative and false positive cases to 5% and 0% respectively. However, the group that we have studied can at most be used as a 'learning set' for deriving a hypothetical predictive function. The latter should be applied prospectively to a large population of IVF trials before deriving firm conclusions (Romeder, 1973). In view of the relatively small number of cases occurring in each centre, this type of study should be carried out on a multicentric basis. References Bardin,C.W. and Lipsett.M.B. (1967) Testosterone and androstenedione blood production rates in normal women and women with idiopathic hirsutism or polycystic ovaries. J. Clin. Invest., 46, 891—902. Blankstein,J., Shalev.J., Saadon.T., Kukia,E.E., Rabinovici.J.. Pariente,C, Lunenfeld,B., Serr,D.M. and Mashiach.S. (1987) Ovarian hyperstimulation syndrome: prediction by number and size of preovulating ovarian follicles. Fertil. Steril., 47, 597 — 602. Cole.H.H. and Hart.G.H. (1930) The potency of blood serum of mare in progressive stage of pregnancy in effecting the sexual maturity of the immature rat. Am. J. Physiol., 93, 57. Devane,G.W., Czekala,M., Judd,H.L. and Yen,S.C. (1975) Circulating gonadotropins, estrogens and androgens in polycystic ovarian disease Am. J. Obstet. GynecoL, 121, 496-500. Dirnfeld.M., Lejeune.B., Camus.M., Vekemans,M. and Leroy.F. (1985) Growth rate of follicular estrogen secretion in relation to the outcome of in vitro fertilization and embryo replacement. Fertil. Steril., 43, 379-384. Figueroa-Cases,P. (1958) Reaccion ovarica monstruosa a las
962
gonadotropines a proposito de un caso fatal. Ann. drug., 23, 116-118. Givens,J.R., Andersen,R.N., Umstot.E.S. and Wiser,W.L. (1976) Clinical findings and hormonal responses in patients with polycystic ovarian disease with normal versus elevated LH levels. Obstet. GynecoL, 47, 338-394. Golan,A., Ron-E1,R., Herman,A., Weimraub,Z., Soffer.Y. and Caspi.E. (1988) Ovarian hyperstimulation syndrome following r>Trp-6 luteinizing hormone-releasing hormone microcapsules and metropin for in vitro fertilization. Fertil. Steril., 50, 912-916. Haning.R.V., Austin,C.W., Carlson,I.H., Kuzma,D.L., Shapiro.S.S. and Zweibel.W.J. (1983) Plasma estradiol is superior to ultrasound and urinary estriol glycuronide as a predictor of ovarian hyperstimulation during induction of ovulation with menotropins. Fertil. Steril., 40, 31-36. Lejeune,B., Degueldre.M., Camus,M., Vekemans,M., Opsomer,L. and Leroy,F. (1986) In vitro fertilization and embryo transfer as related to endogenous luteinizing hormone rise or human chorionic gonadotropin. Fertil. Steril., 45, 377-383. Moses,M., Bogowsky,H., Anteby.E., Lunenfeld.B., Rabau.E., Serr,D., David,A. and Salomy,M. (1965) Thromboembolic phenomena after ovarian stimulation with human menopausal gonadotrophins. Lancet, ii, 1213-1215. Muller,P. (1962) In Beclere.C. (ed.), Gonadotrophines en Gvnecologie. Masson, Paris, p. 137. Orsini,L.F., Venturoh.S., Lurosso.R., Pluchinotta.V., Paradisi,R. and Bovicelli,L. (1985) Ultrasonic findings in polycystic ovarian disease.
Fertil. Steril., 43, 709-714. Puissant,F., Van Rysselberge,M., Barlow,P., Deweze,J. and Leroy.F. (1987) Embryo scoring as a prognostic tool in IVF treatment. Hum. Reprod., 8, 705-708. Rabau,E., David,A., Serr.D.M., Mashiach,S. and Lunenfeld,B. (1967) Human menopausa] gonadotropins for anovulation and sterility. Am. J. Obstet. GynecoL, 98, 92. Romeder J.M. (1973) Mtthode et Programmes d'Analyse Discriminante. Dunod, Paris, p. 66. SchenkerJ.G. and Weinstein.D. (1978) Ovarian hyperstimulation syndrome: a current survey. Fertil. Steril., 30, 255—268. Tal,J., Paz,B., Samberg.I., Lazaroun.N. and Sharf.M. (1985) Ultrasonographic and clinical correlates of metropin versus sequential clomiphene citrate: metropin for induction of ovulation. Fertil. Steril., 44, 342. Tibi,C, Alvarez.S., Comet,D., Antoine,J.M., Gomez,A.C. and SalatBaroux,J. (1989) Prediction des hyperstimulations ovariennes. Contracept. Fertil. Sexual.. 17, 751-752 Received on January 14, 1991; accepted on April 25, 1991
