全文

Objective
The objective of the study was to determine the fecundity of young women (<35 years) with an elevated day 3 follicle stimulating hormone (FSH) undergoing gonadotropin-stimulation/intrauterine insemination.

Study Design
This was a retrospective study. The study was conducted at an academic fertility center. A total of 1396 gonadotropin stimulation/intrauterine insemination cycles from 563 women were stratified by day 3 FSH levels (<10 vs ≥10 U/L) and outcomes were compared. Gonadotropin dose, treatment duration, peak estradiol (E2), number of preovulatory follicles (total, large, and medium size), E2/follicle, endometrial thickness, spontaneous abortion, clinical and multiple pregnancy rates were measured.

The statistics included a Student t test, a χ2, regression, and a discrete survival analysis.

Results
An elevated day 3 FSH was found in 10.2% of the women, despite favorable age (31.9 ± 2.5 years). Women with a day 3 FSH of 10 U/L or greater when compared with women with a normal day 3 level required significantly more medication (1058.9 ± 1106.0 vs 632.7 ± 477.5 IU, P < .0001) were triggered a day earlier (10.6 ± 2.4 vs 11.5 ± 2.9 days, P = .0006) and had E2 levels (on the day of and the day prior to human chorionic gonadotropin administration) that were significantly higher (529.5 ± 244.3 vs 450.0 ± 244.2 and 359.6 ± 141.7 vs 306.8 ± 160.9 pg/mL, respectively, P < .05). Clinical pregnancy rates were comparable among the groups (14.6 vs 14%, respectively, P > .05). Spontaneous abortion and multiple pregnancy rates were higher among women with an FSH of 10U/L or greater but not significantly so (27.8% vs 12.0%, 22.2% vs 13.8% for FSH of ≥10 vs FSH < or >10 U/L, P > .05).

Conclusion
Women younger than 35 years with an elevated day 3 FSH, when treated aggressively with gonadotropins have pregnancy rates comparable with those of women with a normal baseline FSH. To achieve this outcome, they need higher doses of medication to stimulate the production of a larger preovulatory follicular cohort.

Key words: day 3 follicle stimulating hormone; gonadotropins; infertility; intrauterine insemination; ovulation induction


In the setting of infertility, accurate determination of ovarian reserve is extremely important because the information is used by clinicians to determine eligibility and protocols for fertility treatments and to properly counsel couples about the probability of success of the proposed therapies. Several ovarian reserve tests are available and include measurement of day 3 follicle-stimulating hormone (FSH) and estradiol (E2),1, 2, 3 and 4 basal inhibin-B and antimullerian hormone (AMH) levels,5, 6 and 7 the clomiphene citrate (CC) challenge test,2, 8, 9 and 10 and the ultrasonographic evaluation of antral follicle counts (AFC) and ovarian volume.2, 11, 12 and 13 Despite the plethora of available ovarian response biomarkers and the evidence suggesting that AFC and AMH correlate better with true ovarian reserve (histologically confirmed ovarian primordial follicle number) than day 3 FSH does,14 the latter is widely available and often used as either the first-line or the only test in ovarian reserve evaluation.

Elevation in basal FSH levels is thought to reflect ovarian aging and is associated, at least in older women undergoing in vitro fertilization (IVF), with poor ovarian response and low pregnancy rates,4, 15 and 16 with numerous studies also suggesting that decreased ovarian reserve (DOR) is associated with a higher chance of fetal loss17, 18, 19, 20, 21 and 22 and chromosomal abnormalities in the conceptus.23, 24, 25 and 26

Because most of the studies evaluating the association between measures of ovarian reserve and reproductive outcomes have been performed in older patients undergoing aggressive assisted reproduction technology treatments, it is difficult to counsel younger women with elevated basal FSH levels about their appropriate treatment options and expected success rates. Furthermore, basal FSH levels predict more accurately the size of the remaining oocyte pool and not necessarily its quality (the latter declines as age advances and biological damage to the oocyte accumulates).27, 28, 29, 30, 31 and 32 This potentially accounts for the more favorable reproductive outcome of younger poor responders compared with that of their older counterparts.33 and 34

The purpose of the present study was to evaluate the prognostic significance of elevated day 3 FSH levels on the outcome of gonadotropin stimulation (GS)/intrauterine insemination (IUI) cycles in women younger than 35 years. GS/IUI cycles with normal basal FSH levels (<10 U/L) were compared with cycles characterized by elevated FSH levels (≥10 U/L). Thus, we sought to provide the treating physicians with a tool to properly counsel younger patients (with elevated basal FSH levels) regarding less aggressive treatment options such as GS/IUI.

Materials and Methods
The study was approved by the Massachusetts General Hospital Institutional Review Board. All IUI cycles (January 2004 through July 2010) from women younger than 35 years were retrospectively reviewed, and 1396 GS/IUI cycles were identified and further analyzed. Of these, 143 cycles were from women with a basal FSH of 10 U/L or greater.

At the time of treatment initiation, all participants had completed a standard infertility work-up (as described elsewhere).35 Recombinant FSH or human menopausal gonadotropins was started on the third day after a spontaneous or a progesterone-induced menstrual bleed. Response to gonadotropins was monitored and dose adjusted, as needed. Ovulation was triggered with recombinant human chorionic gonadotropin (hCG; Ovidrel; Serono, Norwell, MA), and 35-36 hours later, the IUI was performed.

Once stimulation started, cycles were cancelled either for overresponse (more than 5 preovulatory follicles) or for no response. Pregnancy was defined as clinical when at least 1 gestational sac was visualized on ultrasound 3-4 weeks after the IUI.

Outcome measures included the following: total gonadotropin dose, duration of gonadotropin stimulation, peak E2, number of preovulatory follicles (total, large [≥15 mm], and medium size [13-15 mm]), E2 per follicle produced, spontaneous abortion (SAB), and clinical and multiple pregnancy rates.

Statistics
Means or geometric means between groups were compared using a Student t test or 1-way analysis of variance as appropriate for continuous variables, and χ2, and Fisher exact tests were used to compare categorical data. Multivariable models for associations between day 3 FSH and the outcomes of interest, adjusted for age and body mass index (BMI), were constructed.

For continuous outcomes, linear mixed-effects models were used to account for correlations between cycles within patient. Several variables (gonadotropin dose, peak E2 variables, and peak E2 variables divided by number of follicles) were skewed and transformed by the natural logarithm prior to inclusion in the models. The association of elevated day 3 FSH on the number of large, medium, and total follicles was assessed using generalized estimating equations (GEE) poisson regression models. A logistic GEE model was used to model whether a pregnancy was achieved as well as the odds of a multiple pregnancy or SAB.

Finally, because a live birth occurred only a maximum of 1 time per patient, a discrete survival model that adjusted for the cycle was used to calculate odds ratios for achieving a live birth in relation to elevated day 3 FSH.36 Values of P < .05 were considered statistically significant.

Results
During the study period, 2681 insemination cycles were performed. In 1396 cycles, gonadotropins were used either alone (n = 1391) or in combination with CC (n = 5). Five hundred and two women (89.2%) with a basal FSH less than 10 U/L contributed 1253 cycles (89.8%) and 61 women (10.8%) with an FSH of 10 U/L or greater contributed 143 cycles (10.2%). Overall, GS/IUI cycles represented 57% and 71.8% of the insemination cycles performed among women with normal and elevated basal FSH levels, respectively (P = .0001).

Table 1 summarizes the demographic characteristics stratified by the day 3 FSH. The mean age, day 3 FSH, and BMI of the study participants were 31.9 ± 2.5 years, 7.3 ± 2.7 U/L, and 23.7 ± 4.4 kg/m2, respectively. The age distribution was as follows: 235, 293, and 868 cycles were from women younger than 30 years, 30 years or less to 32 years or less, and between 32 years or less and 35 years or less, respectively.

Women with an elevated day 3 FSH were slightly older than the ones with a normal day 3 value (Table 1) and were less likely to have received CC in the past (14.7 vs 49.1%, respectively, P < .01). The most common infertility diagnosis among the former group was DOR, followed by combined factors, and idiopathic infertility, whereas in the latter, the most prevalent diagnosis was idiopathic infertility, followed by male and combined factors (P < .01 for all comparisons, Table 1).

Women with a day 3 FSH of 10 U/L or greater required significantly more medication and were triggered a day earlier than women with normal day 3 levels. E2 on the day of and the day prior to hCG trigger was significantly higher in women with an elevated day 3 FSH (Table 2).

On average, women with a normal day 3 FSH produced 1.8 ± 0.95 total follicles per cycle, and those with an elevated FSH produced 2.1 ± 0.96 follicles (P = .0004). The percentage of patients who developed 1 follicle only in response to gonadotropins was significantly higher in the normal FSH group with significantly more patients in the high FSH group producing 2-3 follicles (46.9 vs 33.6%; P = .003, and 45.2 vs 59.2%; P = .002, respectively).

Using a GEE model adjusted for age and BMI, we noted that women with an elevated day 3 FSH produced more total follicles than women with a normal value (parameter estimate [PE], 0.14; 95% confidence interval [CI], 0.04–0.24; P = .007) primarily because they produced a higher number of medium size follicles (PE, 0.32; 95% CI, 0.12–0.53; P = .002; Figure 1, A-C). However, using a linear mixed-effects model adjusted for age and BMI, we noted that women with an elevated day 3 FSH produced fewer total follicles per given FSH dose (PE, –0.8; 95% CI, –1.5 to –0.13; P = .02).

With regard to the outcome of the cycle, there was no difference in the number of pregnancies achieved. The clinical pregnancy rate per cycle was 14.6% and 14.0% for FSH less than 10 and FSH 10.0 U/L or greater, respectively (14.1% and 13.8%, respectively, P = 1.0, when limiting the analysis to ovulatory women only). Sixty-nine percent and 72% of the pregnancies were achieved by the second cycle (P = 1.0), whereas 32.4% and 29% of the women attempting to conceive eventually did (P = .7).

Adjusting for age and BMI and using GEE to calculate the odds for pregnancy and discrete survival analysis to calculate the odds for live birth, we noted no difference between the 2 groups (odds ratio [OR], 1.2; 95% CI, 0.6–1.7, P = .9, and OR, 0.9; 95% CI, 0.4–1.8; P = .7, respectively; Figure 2, A). There were more miscarriages and multiple pregnancies among the women with an elevated FSH, but differences were not significant (27.8% vs 12%, P = .074; 22.2% vs 13.8%, P = .3 for FSH of ≥10.0 vs <10 U/L, respectively). However, the former SAB rate was higher than expected for women younger than 35 years. Adjusting for age and BMI and using GEE to calculate the odds for SAB and discrete survival analysis to calculate the odds for multiple pregnancy, we noted no significant difference between the 2 groups (OR, 2.0; 95% CI, 0.5,–7.8; P = .3, and OR, 3.0; 95% CI, 0.7–13.2; P = .15, respectively; Figure 2, B).

Cancellation rates were similar (8.4% vs 9.1% for FSH less than 10 vs FSH 10 U/L or greater, respectively, P = .57), although more cycles were cancelled for overresponse in the former (36.4% vs 0.0%) and for low or no response in the latter (20.0% vs 69.2%, respectively, P = .0005). In the latter, the vast majority of cancellations (92%) occurred in women older than 33 years of age. Furthermore, women with an elevated FSH whose cycles were cancelled were older and heavier than those with a normal FSH and cancelled cycles (33.5 ± 1.6 vs 31.0 ± 2.9 years; P = .003, 26.6 ± 7.5 vs 23.7.0 ± 4.3 kg/m2; P = .037, respectively). In the group of women with an FSH of 10 U/L or greater, cancelled cycles were characterized by a significantly higher basal FSH and BMI than those completed and leading to an IUI (FSH: 15.9 ± 4.7 vs 12.8 ± 3.2 U/L, P = .0016; BMI: 26.6 ± 7.5 vs 22.8 ± 4.4 kg/m2, P = .007).

Day 3 FSH levels did not correlate with either the peak E2 levels or with the E2 level (both on the day of and the day prior to hCG trigger) per produced preovulatory follicle.

Among patients with an elevated day 3 FSH, those who conceived did not differ in age, day 3 FSH, medication dose, total number of follicles, and number of follicles greater than 15 mm from the ones that did not conceive. However, most pregnancies (89%) were seen in women with an FSH less than 14.0 U/L (representing 74.2% of the elevated basal FSH cycles). Furthermore, in the latter subgroup, pregnancy rates were higher and SAB rates lower (16.0% vs 5.4%, and 25.0% vs 50.0%, for FSH less than 14.0 vs 14.0 U/L or greater, respectively), but numbers were too small to reach meaningful statistical conclusions.

Among women with an elevated basal FSH, more cancellations were noted in the subgroup of women with an FSH of14.0 U/L or greater, but the difference did not reach statistical significance (16.2% vs 6.6%, P = .1).

Among patients who conceived, the mean number of cycles per patient required to achieve a clinical pregnancy did not differ significantly between normal (1.95 ± 0.93) and elevated day 3 FSH women (1.94 ± 1.3).

Comment
Decreased ovarian reserve is a rather common problem among subfertile women, with approximately 15% of the cycles reported by the Society of Assisted Reproductive Technology complicated by the sole diagnosis of DOR.37 and 38 One study estimated that DOR affected 10% of the women attending an infertility practice, with an increase in prevalence with advancing age.39 Similarly, in our study DOR prevalence was 10.2%, despite favorable reproductive age.

In general, the DOR diagnosis reflects the process of ovarian senescence and is associated with adverse reproductive outcomes.40 and 41 Women with an elevated day 3 FSH undergoing IVF are at risk for suboptimal response to stimulation, poor oocyte yield, high cancellation rates, and a low chance of success.4, 16 and 42 Furthermore, an association between DOR and increased SAB rates has been reported.21, 22 and 43

To evaluate ovarian reserve and properly counsel couples, physicians take into consideration multiple predictors, including the female's age and the results of ovarian response testing.2, 3, 4, 5, 6, 7, 8, 9, 11 and 13 Despite the advent of new markers (AMH, AFC), day 3 FSH levels are used very frequently (often as the sole marker) for the prediction of ovarian reserve. Because most studies associating elevated basal FSH levels with adverse reproductive outcomes evaluated older patients pursuing IVF following the failure of conservative therapies (ie, ovulation induction/IUI), the question that arises is whether these results are generalizable to younger women at the very early stages of fertility treatments. The current study (designed to investigate the consequences of an elevated basal FSH on the outcome of GS/IUI cycles in women younger than 35 years) found that an elevated day 3 FSH was not associated with a deleterious effect on the outcome of the treatment cycle.

In the current study, 10% of women presenting for evaluation of infertility had an FSH of 10 U/L or greater, despite a favorable reproductive age and treating physicians showed a preference to stimulating them with gonadotropins rather than CC, a finding possibly reflective of the physician's and patient's concerns regarding the risk of a poor response and adverse reproductive outcome.

Also, a significant trend toward higher medication requirements was noted with increasing basal FSH levels. This finding might reflect either the increased gonadotropin needs of the remaining oocyte pool or a possible physician bias in the selection of the gonadotropin dose, possibly arising from the prior knowledge of the elevated basal FSH level. In our study population, young patients with an elevated FSH responded favorably to the increase in the medication dose with higher peak E2 levels and a higher number of recruited follicles. When compared, though, with women with normal FSH values, they produced fewer follicles per given FSH dose, a finding potentially reflecting more the resistance to gonadotropins of the decreasing oocyte pool rather than a possible bias from the treating clinician. The observed trend toward higher medication requirements is in agreement with findings of IVF studies.44

As expected, the cancellation rates for no response were more commonly seen among patients with an elevated FSH, whereas among those with normal FSH levels, the most common reason for cancellation was overresponse. These findings are similar to those seen in IVF patients in whom the cancellation rates are significantly higher among patients with elevated day 3 FSH levels compared with patients with normal values across all age groups.32, 44 and 45 In this aspect, young patients with an elevated basal FSH undergoing GS/IUI do not behave differently from women undergoing IVF because they manifest similar rates of diminished follicular response.

In our study, women with an elevated day 3 FSH were administered hCG at an earlier point than women with a normal value, a finding consistent with the early follicular recruitment often seen in cycles of women with a shrinking oocyte pool. More importantly, pregnancy rates per cycle were comparable among the 2 groups, with two-thirds of the pregnancies achieved by the second cycle in both groups. Approximately one-third of the patients attempting conception through GS/IUI eventually conceived, and this percentage was similar among both groups. The SAB rates, though, were higher among women with an elevated day 3 FSH but not significantly so.

This finding is consistent with that of other authors44 and 46 who showed that in an IVF population the miscarriage rates were affected by age but not by the FSH levels. It also contradicts data suggesting that poor responders have higher miscarriage rates irrespective of age.47The population of the latter study, however, was very heterogeneous, comprised of poor responders with high and normal basal FSH values.

Whereas many studies have reported on the clinical value of basal FSH in the prediction of IVF outcomes, none has reported on its value in relation to the outcome of the GS/IUI cycles. One study31 investigated the predictive value of basal FSH on spontaneous pregnancy in subfertile ovulatory women and found that a basal FSH of greater than 8.0 U/L was associated with decreasing fecundity, independent of female age and cycle length.

To the best of our knowledge, this is the only study investigating the clinical value of an elevated basal FSH level in predicting the outcome of GS/IUI cycles among women of favorable age. Our study analyzed a large number of cycles all performed in 1 fertility center with the use of uniform stimulation protocols. FSH was measured in 1 laboratory, thus minimizing differences arising from assay variations.

A possible limitation of our study is that we recorded only 1 basal FSH level, the one measured at the time of infertility evaluation and prior to the onset of the fertility treatments. Because there are intercycle variations, it is possible that repeat measurements might have identified more patients with basal FSH levels greater than 10 U/L. However, repeated FSH measurements did not provide a substantial benefit in predicting poor ovarian response in IVF cycles.1 Furthermore, in recent years, the measurement of serum AMH levels has been added in the evaluation of ovarian reserve.5, 6 and 7

AMH has been proven to reliably predict the ovarian response to controlled ovarian hyperstimulation and perhaps is a much more predictive measure of ovarian reserve than other modalities.14 However, this assay is not used routinely and was not available at our center during the study period. Therefore, the lack of AMH data might be a limitation of our study. Moreover, given the low incidence of DOR among younger women, our study might have been underpowered for detecting a significant difference in the SAB and multiple pregnancy rates between the 2 groups.

Finally, the findings of our study provide a tool to counsel young women with an elevated basal FSH contemplating fertility treatments. In the current era, there is increasing interest in improving clinical outcomes, increasing treatment efficiency, and decreasing health care costs. In this day and time, it becomes important to correctly identify patients who might benefit from certain testing and therapeutic interventions to correctly allocate the appropriate resources.

In conclusion, day 3 FSH levels are often used as predictors of ovarian reserve, ovarian response to stimulation, and outcome. Elevated FSH levels possibly reflecting a shrinking oocyte pool are not uncommon among younger women evaluated for infertility, and in the current study were found in at least 10% of women and even in women in their early 20s. When counseling young women with DOR, though, one should treat them differently than their older counterparts. Gonadotropin stimulation/IUI should be attempted in these women because cycles not cancelled for low response have a reasonable chance of live birth, despite increased medication requirements. Patients should not be discouraged from attempting treatments because the follicular pool might be decreased but not to a critical point beyond which decreased chances of conception and live birth occur. To this end, it is critical to individualize stimulation with respect to endocrine status and age, not hesitating to increase the dose of the medication to meet the increased medication requirements of the diminishing oocyte pool.