Ovarian stimulation length, number of follicles higher than 17 mm and estradiol on the day of human chorionic gonadotropin administration are risk factors for multiple pregnancy in intrauterine insemination

Abstract

Aim:  The aim of the present study was to identify the risk factors, their prognostic value on multiple pregnancies (MP) prediction and their thresholds in women undergoing controlled ovarian hyperstimulation (COH) with follicle stimulating hormone (FSH) and intrauterine insemination (IUI).

Methods:  A case‐control study was carried out by identifying in our database all the pregnancies reached by donor and conjugal IUI (DIUI and CIUI, respectively), and compared cycle features, patients’ characteristics and sperm analysis results between women achieving single pregnancy (SP) versus MP. The number of gestational sacs, follicular sizes and estradiol levels on the human chorionic gonadotropin (hCG) administration day, COH length and semen parameters were obtained from each cycle and compared. Student's t‐tests for mean comparisons, receiver–operator curve (ROC) analysis to determine the predictive value of each parameter on MP achievement and multiple regression analysis to determine single parameter influence were carried out.

Results:  Women with MP in IUI stimulated cycles reached the adequate size of the dominant follicle (17 mm) significantly earlier than those achieving SP. Also, the mean follicles number, and estradiol levels on the hCG day were higher in the CIUI and DIUI MP group. Nevertheless, only ROC curve analysis revealed good prognostic value for estradiol and follicles higher than 17 mm. Multiple regression analysis confirmed these results. No feature of the basic sperm analysis, either in the ejaculate or in the prepared sample, was different or predictive of MP. When using donor sperm, different thresholds of follicle number, stimulation length and estradiol in the prediction of MP were noted, in comparison with CIUI.

Conclusions:  MP in stimulated IUI cycles are closely associated to stimulation length, number of developed follicles higher than 17 mm on the day of hCG administration and estradiol levels. Also, estradiol has a good predictive value over MP in IUI stimulated cycles. The establishment of clinical thresholds will certainly help in the management of these couples to avoid undesired multiple pregnancies by canceling cycles or converting them into in vitro fertilization procedures. (Reprod Med Biol 2007; 6: 19–26)

INTRODUCTION

INTRAUTERINE INSEMINATION (IUI) with either husband's or donor sperm has for decades been the initial and most commonly used assisted reproduction technique for infertile couples. ¹ Until the advent of in vitro fertilization (IVF) in 1978 and intracytoplasmic sperm injection (ICSI) in 1992, IUI was the only alternative for infertile couples who were unsuccessfully trying to conceive by sexual intercourse.

The mechanisms by which IUI is increasing the pregnancy possibilities are assumed to be: (i) an increase in the number of selected spermatozoa at the site of fertilization; ² (ii) adequate ovulation timing; ³ (iii) the bypass of hostile cervical mucus or anatomical difficulties in the cervix; (iv) the enhancement of multiple follicular development; and (v) women's exposure to prepared sperm in the exact period of the ovum availability are surely decisive for success. ⁴ Finally, sperm preparation mimicking the natural capacitation process is needed to remove prostaglandins, antigenic molecules and infectious agents in the seminal plasma, and select the best motility and shaped spermatozoa, discarding immature sperm cells, immotile spermatozoa and immune cells within the ejaculate.

Conjugal IUI (CIUI) is appropriated in non‐severe male factor infertility or impossibility to deposit the ejaculate in the vagina (hipospadias, impotence, retrograde ejaculation, vaginism etc.), women with anovulation, polycystic ovarian syndrome and mainly unknown origin infertility when no abnormalities can be confirmed after infertility work‐up. ¹ Conversely, donor IUI (DIUI) is mainly reserved for couples with severe male factor infertility, such as severe oligo‐asthenozoospermia (rejecting IVF/ICSI or after these treatment's failure) or azoospermia, couples in which the male is at risk of transmitting a genetic or infectious disease, or women without a male partner (homo or heterosexual) who desire maternity. ⁴

In our experience, the minimum requirements to expect optimum pregnancy rates are 2 000 000 totally motile sperm being inseminated over two consecutive days and healthy women with tubal permeability undergoing controlled ovarian hyperstimulation (COH) cycles. ⁴ The mean expected results are 20% and 25% per cycle with 60–70% and 75–80% of the women becoming pregnant in four consecutive treatments of CIUI and DIUI, respectively. Nevertheless, several studies have reported diverse success rates, probably as a result of the used of different COH regimes. ⁵ These results have shown to be more cost‐effective than IVF when IUI treatment is adequately prescribed. ⁶

Few complications are associated with these treatments. Allergic reactions to some sperm preparation media components such as bovine serum albumin or antibiotics and seminal plasma, together with pelvic infections caused by the involuntary transfer of pathogens are extremely rare (0–0.2%). ⁷

Since the advent of COH, ovarian hyperstimulation syndrome in some cases, but mainly multiple pregnancies (MP) have emerged as the most common iatrogenic problems. ⁸

The rise in MP is linked to the increased incidence of assisted reproduction treatments in the past 20 years, mostly IUI rather than IVF. In these cases, perinatal morbidity and mortality are increased, together with early and late neonatal deaths, and both low birthweight and prematurity are increased with the number of fetuses. ⁹

Furthermore, unique embryo transfer aiming for the reduction of MP has been a major concern in several countries recently, and a great deal of effort has been dedicated to this issue. The different reproduction societies ⁹ , ¹⁰ have supported the selection of the optimal embryo to be transferred, so maximum pregnancy probabilities are guaranteed, thus aiming to decrease the number of low and high order multiple pregnancies. ¹¹

To our knowledge, there are few works available in the literature regarding IUI management and the identification of factors involved in achieving MP so as to develop new strategies to diminish the probability of multiple conceptions.

Nevertheless, several studies have attempted to investigate these issues, but in their design they included women in which IUI was unsuccessful (not reaching pregnancy), thus masking the real problem; the differences exhibited by those women pregnant with one sac versus more than one sac. ¹²

Gleicher et al. stated that less intensive ovarian stimulation regimes could help in the reduction of the incidence of MP, but ‘only to a limited extent and at expense of overall rates’. ¹³ Also, they considered the number of follicles and estradiol levels common cut‐off values not to be associated with MP.

Then, it is interesting to determine the risk factors to achieve MP in the treatments while the best pregnancy chances are maintained (those with COH with follicle stimulating hormone [FSH] as previously described). However, no thresholds from which a MP risk is increased have been defined yet. It would be undoubtedly useful in the management and counseling of couples undergoing IUI, because COH at high risk of MP could be cancelled in order to prevent multiple gestations, or could be transformed into an IVF cycle. Nevertheless, to maintain the best pregnancy chances for our patients, the use of COH with FSH is mandatory.

Subsequently, in the present study, we aimed to carry out a case–control study by identifying MP in our database, analyzing the patients’ cycle characteristics after COH with FSH followed by IUI, and comparing their results with a single pregnancy (SP) control group. The aim was to identify risk factors and their prognostic value which predispose patients to MP, in order to help in the routine management of IUI cycles.

MATERIALS AND METHODS

THE PRESENT WORK was supervised and approved by the Institutional Review Board on the use of human subjects in research at the Instituto Universitario IVI, and complies with the Spanish Law of Assisted Reproductive Technologies (35/1988).

The study was designed as a retrospective cases and controls analysis, given the low incidence of multiple gestations. We identified those women undergoing IUI with either donor or husband's sperm, who achieved MP. Then, we searched for a control group in the same setting where SP was achieved. Furthermore, the sample size calculation showed that the number of patients included in the present study was sufficient to statistically confirm minimum differences between groups. ¹⁴

Subsequently, we obtained the data from our database; 1362 DIUI and 2851 CIUI, where we identified 340 (24.9% per cycle) and 570 (19.9% per cycle) pregnancies with ecographically confirmed sacs, respectively, from January 1998 until December 2005. A total number of 448 and 580 patients undergoing DIUI and CIUI, respectively, were eligible for the study. Miscarriage rate was 15.4% and 19.0% in the DIUI and CIUI groups, respectively.

No ovarian hyperstimulation syndrome was observed in our IUI patients, mainly as a result of the soft stimulation protocols used.

Women's mean age was 32.6 ± 3.6 years, mean stimulation length was 7.8 ± 2.7 days, and mean number of follicles and estradiol on the hCG day was 1.9 ± 1.4 and 947 ± 704 pg/mL, respectively.

We only included those IUI cycles stimulated with FSH, which actually is the most frequent protocol in our setting. Only inseminations reaching 2 000 000 of totally motile sperm in two consecutive days were included in order to standardize the male factor. Other stimulation regimes, as well as non‐stimulated cycles, were excluded from the study.

All women had a previous investigation of tubal permeability and ovarian function as determined by day 3 hormone determination (FSH, luteinizing hormone [LH] and estradiol).

Semen analysis

Semen parameters were evaluated for every ejaculate by two independent observers. After the liquefaction of the semen at 37°C, 5% CO₂, for 10 min, the samples were examined for concentration and motility according to the WHO guidelines ¹⁵ in a Mackler chamber (Sefi Laboratories, Tel Aviv, Israel). Sperm morphology was studied following Tygerberg's strict criteria. ¹⁶ Briefly, spermatozoa were divided between normal and abnormal, and the latter were also categorized depending on the type of defect; head defect, midpiece defect and tail defect, and their respective percentages were recorded. Results were only accepted when the differences between two observers were less than 5%.

Sperm preparation in conjugal intrauterine insemination

All samples for CIUI were always prepared by swim‐up technique. Briefly, raw ejaculates were diluted 1:1 (vol : vol) with Sperm Medium (MediCult, Jyllinge, Denmark). Then, they were pelleted at 400 g for 10 min and the supernatants were discarded. This was followed by the careful addition, without disturbing the pellet, of 0.5–1 mL fresh medium and the incubation for 45 min of the tubes with a 45° inclination.

After the sperm preparation, 0.5 mL were taken and analyzed before the transference of the prepared sperm into the uterus.

Sperm preparation in donor intrauterine insemination

For the insemination, frozen sperm was always used in DIUI, after their corresponding quarantine that allowed the confirmation of the absence of infectious diseases in the donor (confirming this fact with a blood analysis at least 6 months after the sample donation and freezing).

Samples were subsequently thawed when needed after their selection to match the receiver couple's phenotypical characteristics as well as blood type, as described in the Spanish Assisted Reproduction Law. Then, samples were prepared by density gradients, as described elsewhere. ⁴

Controlled ovarian hyperstimulation

Briefly, after confirming the absence of ovarian cysts by ultrasonography, ovarian stimulation was carried out by a standard starting dose depending on the patient's characteristics, age and body mass index, previous cycles, and ovarian response was controlled by both serum estradiol determinations and transvaginal ultrasound scan.

The criterion for the hCG administration was the presentation of a leading follicle with the diameter = 17 mm. Two consecutive intrauterine inseminations were carried out approximately 12 and 36 h after the 10 000 IU hCG administration. Luteal phase support was carried out with 200 mg/day of progesterone until the pregnancy test (Progeffik, Effik Laboratories, Madrid, Spain).

Pregnancy tests were carried out on day 14 after the first insemination by the quantification of beta fraction of hCG and then confirmed 10 days later by vaginal ultrasonography. Those pregnancies spontaneously interrupted before the 20th week of gestation after previous scan detection of the embryo's heartbeat were considered as spontaneous abortions.

We compared overall cycle characteristics (stimulation protocol and length, female etiology, age, estradiol, number of follicles and sperm features) between women who achieved pregnancy with one fetus (SP) versus. multiple fetuses with heartbeat (MP) in order to determine those factors predisposing to multiple pregnancies in both husbands and DIUI.

We included CIUI and DIUI in the present study because previous reports have shown differences in the results obtained. Men belonging to infertile couples, even with comparable basic sperm analysis results, have been shown to present several occult differences in some fertility markers in comparison to sperm donors, such as oxidative stress implicated molecules, calcium, cholesterol, mitochondrial activity and sperm DNA fragmentation. ¹⁷ , ¹⁸

Thus, these factors are possibly implicated in fertilization, implantation and pregnancy, and could predispose differentially to multiple pregnancies between CIUI and DIUI.

Statistical analysis

Statistical analysis was carried out by Student's t‐test for mean comparisons, receiver–operator curves (ROC) to determine the prognostic values of each parameter and multiple regression analysis to study the relationship between the number of sacs and all significantly different variables between groups to weigh the single relevance of each.

χ²‐tests were carried out in contingency tables to compare MP rates between groups depending on the stimulation length, number of follicles and estradiol levels on the day of hCG administration.

Statistical analysis was carried out using the Statistical Package for the Social Sciences (SPSS, Chicago, IL, USA) and MedCalc Software (Ghent, Belgium).

RESULTS

IN THE STUDY population, MP rates were 25.0% and 23.6% in DIUI and CIUI, respectively, without significant differences between groups.

We hypothesized that women's exposure to sperm samples with different quality could be translated into different fertilization, implantation and gestation possibilities. This means high quality samples would present higher probabilities of fertilizing more than one oocyte, if available. To address this issue we compared sperm density, motility, percentage of normal spermatozoa and the different indexes resulting from their combination; total motile progressive (TMP) and total motile progressive with normal morphology (TMPN) in both fresh and capacitated samples between those couples achieving single versus multiple pregnancies. No parameter analyzed was different between them.

As shown in Table 1, we found no differences between the above mentioned sperm features before and after preparation, in neither CIUI nor DIUI regardless of whether SP or MP was reached. This was further confirmed after the ROC curve analysis of the probability of achieving a SP or MP was used to determine the predictive value of sperm characteristics on multiple pregnancies. No area under the curve exceeded a 0.55 value (data not shown), thus showing the lack of predictive value.

Table 1.

Male factor parameters in women achieving single versus multiple pregnancies in conjugal intrauterine insemination and donor intrauterine insemination

	Parameter	Single pregnancy			Multiple pregnancy			P‐value
		Mean (± SD)	Median	Range	Mean (± SD)	Median	Range
CIUI	Volume	3.7 ± 1.8	3.2	0.5–11	3.5 ± 1.7	3.2	0.1–9	0.73
	Concentration	60.2 ± 39.6	53.6	2.8–263	56.3 ± 34.0	54.5	1.6–170	0.89
	A + B%	46.1 ± 14.7	47.0	3–80	48.7 ± 15.6	49	12–83	0.20
	NSR	8.5 ± 4.8	8	0–25	9.0 ± 4.9	8	1–21	0.66
	TMP	14.8 ± 12.1	12.2	0.7–72	13.3 ± 10.7	10	2–22	0.29
	TMPN	3.7 ± 2.1	4.5	0–20.3	3.4 ± 2.6	5.1	0–25.6	0.32
	Concentration post	37.1 ± 28.6	32	1.6–160	11.1 ± 5.1	26	0.8–156	0.31
	A + B% post	77.9 ± 15.3	83.3	14–100	79.5 ± 16.5	79.3	20–98	0.47
	NSC	10.6 ± 5.6	9.5	2–30	11.5 ± 5.1	10.0	2–22	0.60
	TMPI	13.7 ± 10.2	12.7	0.4–40.2	12.6 ± 8.9	13.0	0.6–35.6	0.61
	TMPNI	1.2 ± 0.8	6.6	0–8.3	1.2 ± 0.9	5.9	0–6.9	0.26
DIUI	Volume	2.2 ± 1.5	1.9	0.5–7	1.8 ± 1.3	1.5	0.5–8.4	0.78
	Concentration	47.6 ± 12.3	42.1	5–270	52.1 ± 31.6	50	8–120	0.56
	A + B%	33.6 ± 14.5	31	5–75	35.6 ± 12.0	35	5–70	0.20
	TMP	28.4 ± 15.3	23.2	0.1–161	30.1 ± 16.3	28	1–18	0.80
	Concentration post	21.6 ± 12.6	18	1–98	22.6 ± 11.9	18	5–60	0.74
	A + B% post	57.8 ± 20.0	57	14–97	59.5 ± 19.7	59	24–98	0.77
	TMPI	5.6 ± 3.3	4.1	0.1–43.2	6.0 ± 3.5	4.8	2–24	0.51

A + B%, sperm with progressive motility; CIUI, conjugal intrauterine insemination; DIUI, donor intrauterine insemination; NSC, normal spermatozoa after capacitation; NSR, normal spermatozoa in the raw ejaculate, TMP, total motile progressive spermatozoa in the ejaculate; TMPI, total motile progressive spermatozoa inseminated; TMPN, total motile progressive spermatozoa normal in the ejaculate; TMPNI, total motile progressive spermatozoa normal inseminated.

When female characteristics were analyzed, significant findings were obtained. These findings were mainly related to stimulation length, the number of mature follicles and estradiol levels on the hCG administration day, as shown in Table 2.

Table 2.

Female factor parameters in women achieving single versus multiple pregnancies in conjugal intrauterine insemination and donor intrauterine insemination

	Parameter	Single pregnancy			Multiple pregnancy			P‐value
		Mean (± SD)	Median	Range	Mean (± SD)	Median	Range
CIUI	Cycle no.	1.8 ± 1.0	2	1–5	2.1 ± 1.2	2	1–6	0.59
	Days of stimulation	7.8 ± 2.3	7	1–25	8.5 ± 2.4	8	1–23	0.049
	Follicles >17 mm	1.9 ± 1.0	1	0–9	2.5 ± 1.1	2	0–9	0.001
	Age (years)	32.3 ± 3.1	32	24–42	32.1 ± 3.3	32	21–39	0.52
	Estradiol	860 ± 660	753	132–5148	1553 ± 1002	1430	322–4435	0.0001
DIUI	Cycle no.	2.3 ± 1.6	2	1–8	2.1 ± 1.4	2	1–6	0.69
	Days of stimulation	7.1 ± 2.9	7	1–21	8.3 ± 2.8	7	5–21	0.040
	Follicles >17 mm	1.8 ± 1.1	1	1–6	2.2 ± 1.4	2	1–9	0.047
	Age (years)	34.4 ± 4.0	33	23–43	33.6 ± 3.5	33	25–42	0.55
	Estradiol	712 ± 574	561	110–3231	1027 ± 578	790	418–2840	0.006

CIUI, conjugal intrauterine insemination; DIUI, donor intrauterine insemination.

The present results indicate that in women achieving MP by IUI with COH, the dominant follicle reached the adequate size (17–18 mm) later than those achieving single pregnancies (7.9 vs 8.5 days in CIUI and 7.1 vs 8.3 days in DIUI). Also, mean number of follicles was higher in MP group (1.9 vs 2.5 follicles, P < 0.001 in CIUI and 1.8 vs 2.2, P < 0.05 in DIUI) and serum estradiol levels on the hCG day was significantly higher in MP when compared with SP in donor sperm (712 vs 1027 pg/mL), and (866 vs 1553 pg/mL).

ROC curve analysis showed good prognostic value for estradiol determinations (area under the curve: 0.74 for CIUI and 0.7 for DIUI) followed by number of follicles higher than 17 mm (0.65 for conjugal and donor) (Table 3).

Table 3.

Receiver–operator curve analysis for the predictive value of cycle parameters on multiple pregnancies

	Parameter	AUCROC (95% CI)	Threshold	Sensitivity (%)	Specificity (%)	PPV	NPV
CIUI	Days of stimulation	0.57	7	65.7	54.6	27.7	79.8
	Follicles >17 mm	0.65	2	40.7	78.6	37.7	81.0
	Age (years)	0.52	33	36.7	70.7	80.3	25.5
	Estradiol	0.75	1300	57.6	88.2	54.3	87.7
DIUI	Days of stimulation	0.58	6	82.7	32.9	29.3	85.0
	Follicles >17 mm	0.59	1	63.5	56.5	33.0	82.1
	Age (years)	0.60	32	70.2	56.1	35.0	80.2
	Estradiol	0.71	408	100	39.6	43.9	100

AUC, area under the curve; CIUI, conjugal intrauterine insemination; DIUI, donor intrauterine insemination; NPV, negative prognostic value; PPV, positive prognostic value, ROC, receiver–operator curve.

To discern the relevance of each single parameter, we carried out multiple regression analysis of those that showed significantly different results.

Multiple regression analysis of significant variables confirmed the relevance of estradiol determinations on the hCG days, higher than the relevance of stimulation length or follicles higher than 17 mm on the ovulation induction day (coeff. 0.00021, P < 0.02; F‐Ratio 4.6, P = 0.005 for DIUI, and coeff. 0.00018, P < 0.0003, F‐Ratio 6.9, P < 0.001) for CIUI, respectively.

Then, we compared the MP rates depending on the different number of follicles available on the hCG day, as well as the estradiol level on this day and the stimulation length. The results can be found in Figure 1.

Figure 1.

Multiple pregnancy rates depending on the number of follicles and estradiol on the human chorionic gonadotropin days, and stimulation length as grouping variables. All superscripts indicate P < 0.05 in comparison with the remaining columns within the grouping variables. Those duplicated superscripts indicate P < 0.05 between marked columns. For the comparisons, χ²‐tests were carried out. (□) conjugal intrauterine insemination; (▪) donor intrauterine insemination; () combined intrauterine insemination.

Basically, it is confirmed that developing one follicle of at least 17 mm on the hCG day during COH is, as expected, linked to the lowest MP rates. Increasing the follicle number to two is statistically enhancing the MP risk to nearly 37% per pregnancy by IUI. These rates are progressively being raised as the number of mature follicles increase, in both CIUI and DIUI.

Taking into account the length of the stimulation, those women with a dominant follicle reaching the adequate size in six or fewer days have significantly lower MP rates than those who reach the adequate size in seven or more.

Estradiol levels lower than <1400 pg/mL showed lower MP rates than those higher than 1400 pg/mL in CIUI, at least fourfold. Whereas, in DIUI this difference is only found in women with less than 600 pg/mL of estradiol in comparison with the remaining groups.

DISCUSSION

SEVERAL EFFORTS HAVE been made in recent years to reduce MP rates, which are a side‐effect of assisted reproduction treatments. Many working groups have been constituted to devise guidelines on this issue, mainly in an attempt to achieve a single embryo pregnancy policy. ¹⁰ , ¹⁹

Furthermore, embryologists are increasing the number of published papers dedicated to determine the embryo quality parameters with maximum probabilities to implant. ¹¹

Another great source of MP linked to assisted reproduction treatments is IUI. Up to 20–25% of the pregnancies achieved by IUI are multiple, ⁸ , ¹² , ¹⁹ which is comparable to the results of IVF treatment. Approximately 5% of MP are triplets or more. Several problems and complications arise in these pregnancies. ⁹

We must assume that our program exhibits quite good pregnancy rates, but also presents a significantly high number of MP; approximately 20% of pregnancies.

To avoid these complications, it is indispensable to know which factors are predisposing to MP in IUI treatments.

Enhancing follicular development has been shown to be an easy and effective way to improve pregnancy rates. Nevertheless, this often leads to MP. To find equilibrium between the improvement of pregnancy rates and avoiding MP, we must know which factors can help us to predict more than one embryo implantation.

Then, if we aim to maintain success rates, we must retain those techniques that were previously shown to positively affect the results, such as FSH induced COH, while changing others to avoid MP. Natural cycles or other stimulation regimes could reduce MP rates, but pregnancy rates could be significantly diminished.

Subsequently, the rationale of the present study is to determine which factors indicate a high probability of having MP in IUI cycles stimulated with FSH.

From the present results, we can conclude that there are three female aspects that must be considered when carrying out IUI, while no male factor is relevant. The stimulation length, the higher number of follicles and estradiol levels on the day of the hCG are the key points in MP, because they are significantly different in women achieving a single or multiple gestation.

Nevertheless, the predictive value of a test used to determine the risk of having a multiple pregnancy is given by the ROC curves analysis. From our results, we can conclude that only estradiol can be efficiently used as a multiple pregnancy predictor.

Multiple comparisons were carried out to determine the relevance of single parameters on the results. Only the estradiol levels measured on the hCG day are independently predictive of other factors.

Several aspects made the present study original and different from the few reports previously available. One of the previous reports ⁸ found that estradiol and the number of follicles higher than 10 mm on the hCG days were predictive factors of MP, together with age. They considered both DIUI and CIUI together in their results and it is obvious that different results obtained when using donor or husband's sperm are reached. There are also noticeable differences between women belonging to one or the other group. ⁴

Differences found between DIUI and CIUI groups are quite surprising and it is difficult to find an explanation for this findings. The different thresholds found for DIUI or CIUI can be explained by the fact that DIUI comprises women with a known and determined cause of infertility; a severe male factor. in these patients, the replacement of the spermatozoa results in an enhanced possibility of reaching pregnancy (as the pregnancy rates show), given that they can be considered in many cases as ‘fertile’.

In contrast, in CIUI, women usually belong to an infertile couple, where the male factor is, at least, non‐severe. These women are clearly infertile, having difficulties in reaching pregnancy. This fact could explain the lower thresholds presented by women undergoing DIUI to be at risk of MP.

This findings also indicate to treat these patients in a different way to those undergoing CIUI.

Using frozen sperm in DIUI, as opposed to fresh ejaculates in CIUI, seems nto be irrelevant, because we can hypothesize only a negative impact of these procedures on the results and this fact is not confirmed.

Similar results regarding the predictive values of age and follicles lower than 10 mm were obtained by Dickey et al., ²⁰ but in that study only low order (singletons and twins) against high order pregnancies were compared.

Nevertheless, estradiol cut‐off point is similar to an intermediate cut‐off considering the data of the present study. Women undergoing DIUI showed a lower cut‐off point regarding estradiol levels than women undergoing CIUI. This could be explained by the DIUI group comprising almost ‘pure’ male infertility within the couple, whereas women are not initially infertile and they seem to be prone to pregnancy easier than women in the CIUI program that belong to an infertile couple. This ‘increased fertility’ is reflected by the ease of achieving pregnancy and MP, thus reducing values of the parameters considered. Together with estradiol, they also achieve pregnancies with fewer days of stimulation and number of follicles.

Regarding age ⁸ we were unable to find any correlation with MP predisposition, although this finding could be related to ovarian reserve and readiness to develop a higher number of follicles in younger women.

The other similar work in the literature ¹² concluded that even 15 mm follicles can initiate a pregnancy. Their design was different to what we presented hereby, because the study only concentrated on follicular size. Also, their pregnancy rates were low (14%) and comprised only 22 MP.

Furthermore, they should take into account whether donor's or husband's sperm is used, because the female population shows differences in the probabilities of having MP. To our knowledge, the finding of the present study are unique. It is well assumed that IUI pregnancy rates achieved by donor sperm are higher than those obtained by husband's sperm. Nevertheless, no previous data was provided regarding MP.

Then, we can conclude that the control of estradiol levels is useful to avoid MP and their associated risks.

This will undoubtedly help in the managment of these couples and assist their physicians in avoiding unwanted MP, and physicians together with the patients could cancel cycles where estradiol levels are over the stated thresholds to avoid unwanted MP.

Even so, to confirm the present results, a prospective trial on the reduction of MP in IUI is needed and is actually being carried out by our group by deciding whether or not to cancel the treatment, taking as a reference the ovarian stimulation length and estradiol on the day of hCG administration.