Objective

To demonstrate that reduction of the cumulative dropout rate (CDR) improves the accuracy of the estimate of the cumulative pregnancy rate (CPR) in a set of four intrauterine insemination (IUI) cycles (“four-cycle program”) and increases the total number of pregnancies obtained.

 

Study design

Single-centre retrospective observational cohort study of couples who underwent IUI cycles at the Andros Day Surgery Clinic, Palermo, from 1997 to 2011. The main outcome measure was the calculation of the CPR, with life table analysis, firstly by giving the same probability of pregnancy to the dropouts as the patients who continued the treatment (usual method) and secondly by considering this probability null (conservative method). The difference between these two methods was used to verify the accuracy of the estimate.

 

Result(s)

In the 15 years, 924 couples underwent 2956 cycles carried out consecutively in a set of four cycles. The CDR was 16%. The CPR was 31.4% with the usual method and 29.1% with the conservative method. The difference between the two estimates was not significant, indicating a high reliability of the results and a good accuracy of the calculation. Furthermore, maintenance of a low CDR permits improvement of the CPR, as was demonstrated by considering scenarios with worse dropout rates.

 

Conclusion(s)

The “four-cycle program” results in a reduction in the CDR, allowing a better estimation of the CPR, and increases the number of actual pregnancies in IUI. The CPR should become the focus for reporting outcome rates in IUI cycles. Reduction of the dropout rate allows us to give the patient a more reliable and accurate estimate of the pregnancy rate.

 

Keywords

Intrauterine insemination; Cumulative dropout rate; Cumulative pregnancy rate; Life table analysis
 

1. Introduction

Intrauterine insemination (IUI) represents the first-line treatment for many forms of infertility, such as unexplained infertility, mild male factor infertility and minimal or mild endometriosis. A debatable topic is which parameter should be used during infertility counselling to provide the most reliable results for patients and physicians [1] and [2]. Because IUI is a simple and repeatable technique, the cumulative pregnancy rate (CPR) should become the reference point for reporting outcome rates.

 

The current methodology for calculating the CPR is life table analysis. Due to the fact that patients who drop out are generally considered to have the same probability of achieving pregnancy as the patients who continue the treatment, this calculation often leads to an overestimation of the CPR [3] and [4]. Another method for evaluating treatment efficacy is to take into account the censored patients. This means that all dropouts are considered to have no chance of conception in subsequent cycles. This conservative method results in underestimation of the CPR.

 

It is known that dropout rates are particularly high among couples with infertility, despite their high levels of motivation [1]. In IUI cycles, Custers et al. [3] found a cumulative dropout rate (CDR) of 28%, and previous studies found CDRs of 37% [5] and 47% [6]. It is notable that when dropout rates decrease, the selection of the method for calculating CPR becomes irrelevant. The reduction in the difference between the traditionally used and conservative approaches may allow better counselling about reproductive prognosis. Furthermore, a significant reduction of dropouts could permit an increase in the number of total pregnancies, with a clinical advantage for couples.

 

In order to achieve dropout reduction, we proposed a new approach in 1997, the so-called “four-cycle program”, instead of the classic cycle-by-cycle program. In the “four-cycle program”, the couple is committed to complete four cycles of treatment, unless pregnancy occurs before the end of the scheduled cycles. The protocol requires full payment in advance for all cycles. The reasons for limiting the number of cycles to four were based on the data available in the literature at the beginning of our experience [7], [8] and [9]. A more recent paper, however, demonstrated that the ongoing pregnancy rates were acceptable up to the ninth cycle [10]. Prior to this fifteen-year experience (data not shown) we had observed a dramatically high CDR when more than four cycles were suggested. In these cases, the positive effect of the pre-payment policy was much less incisive.

 

This paper analyses the IUI cycles performed at the Andros Day Surgery Clinic, Palermo, Italy, from 1997 to 2011. The objectives were to (1) demonstrate that reduction of the CDR improves the accuracy of estimates for the CPR in a set of four intrauterine insemination cycles and (2) verify that the reduction in the CDR increases the total number of obtained pregnancies.

 

2. Materials and methods

In this single-centre retrospective observational cohort study, 984 couples who underwent IUI cycles from 1997 to 2011 were recruited. Before admission to IUI treatment, all couples with infertility lasting for at least 12 months underwent a fertility diagnostic work-up consisting of: clinical history, Chlamydia trachomatis serological titres, tubal patency assessment by hysterosalpingography or laparoscopy with dye test or by transvaginal hydrolaparoscopy with dye test, estimation of mid-luteal progesterone, and semen analysis.

 

Indications for IUI were: unexplained infertility, mild male factor infertility, stage I–II endometriosis according to the ASRM revised classification, anovulation associated with male factor, and other less frequent causes (female sexual pain disorders, and unilateral tubal occlusion). Cervical infertility was not considered because the post-coital test is not part of our diagnostic work-up. Mild male factor was defined according to World Health Organization criteria [11], [12] and [13]. Concerning male factor infertility, couples were considered eligible for IUI if the total number of motile spermatozoa recovered after a preliminary pellet/swim-up test was ≥5,000,000/ml.

 

Two different programmes were proposed to the eligible couples: (i) the “cycle-by-cycle” approach, in which the couples who did not achieve the pregnancy after the first IUI cycle could decide to go on with another cycle and so on until four cycles; (ii) the “four-cycle program”, as described above. This last option was encouraged with a financial advantage (the cost of the “four-cycle program” was 40% less than the amount of four single-cycle costs) in order to minimize the dropout risk for reasons other than clinical.

 

In selecting one of the two approaches, it was clarified that when a pregnancy occurred before the end of the fourth cycle, payment was correlated with this result and no reimbursement of expenditure was envisaged. That is to say, it is likely for there to be two different agreements with the same couple: one for the four-cycle program and the other for having obtained a pregnancy.

 

Specific psychological and medical counselling was provided to give the couples all the necessary information allowing them to make an informed choice and in order to decrease all treatment-related burden that may contribute to treatment discontinuation. During this integrated counselling, the couples had the possibility to express their concerns and to have their misunderstandings addressed. After the counselling, an informed written consent was filled out by the patients.

 

During the 15 years of our experience, the great majority of the couples chose the four-cycle program (n = 924) in comparison with the small number of patients (n = 60) who chose the cycle-by-cycle approach. We collected the data coming only from the 924 couples who were included in the final analysis.

 

Female age, duration of infertility, diagnosis, primary or secondary infertility, the drugs used, and the clinical and multiple pregnancies (divided into twins and high order multiple pregnancies (HOMP)) were registered in a dedicated database. Clinical pregnancies were defined as gestational sacs with a clear-cut fetal heartbeat at 7th week of gestation.

 

The management of patients over time was uniform due to the fact that each couple, in the set of four cycles, was treated by the same physician and with the same protocol. The primary outcome of this study was the difference between the traditional and conservative estimation. The secondary outcome was the increase in the CPR due to the reduction of dropouts. The study was approved by Institutional Review Board of Andros Clinic.

 

2.1. Participants

The female mean age at the time of the first IUI cycle was 32.2 ± 5.1 years. The duration of infertility was 3.53 ± 2.6 years. The infertility causes were: unexplained (484 out of 924 patients, 52.4%), mild or moderate male factor (356 patients, 38.5%), endometriosis stage I–II (31 patients, 3.4%), anovulation associated with male factor (36 patients, 3.9%), unilateral tubal occlusion (14 patients, 1.5%) and female sexual pain disorders (3 patients, 0.3%). The majority of couples suffered from primary infertility (722/924 = 78.1%).

 

Almost all the patients underwent ovarian stimulation (917 out of 924, 99.2%), the majority (870 out of 917, 94.8%) with gonadotropins (50 or 75 or 100 IU/day as starting dose). Different drugs (clomiphene citrate, human menopausal gonadotrophin, urofollitropin, recombinant follicle stimulating hormone, recombinant luteinizing hormone) were used and stimulation protocols were changed for different periods throughout the study.

 

2.2. Statistical analysis

The CPR was calculated in two different ways: firstly, by assigning the dropouts the same probability of pregnancy as the patients who continued the treatment (traditional life table analysis estimation; CPR(i)), and secondly, by considering as null the probability of pregnancy of patients who dropped out (conservative estimation; CPR(ii)).

 

The CDR was defined as the total number of patients who dropped out over the number of patients who continued the treatment.

 

Odds ratio (OR) per cycle in comparison with the first treatment cycle, 95% confidence interval (95% CI) and corresponding p-value for these data, were estimated by regression analysis. Univariate and multivariate logistic regression analyses were performed to estimate the possible predictive effect of prognostic variables on pregnancy rate (PR). Distributions between groups were compared with Pearson's chi-square test. p-value <0.05 was considered to be statistically significant. Statistical analysis was performed using PASW 17.0 (SPSS, Chicago).

 

3. Results

In 15 years (1997–2011), 924 couples underwent a total of 3750 cycles. Of these, 2956 cycles were performed consecutively in a set of four cycles, and 794 cycles subsequently after the set of four cycles and carried out by the couples who did not achieve a pregnancy during the scheduled treatment and who turned down the IVF option. The 794 cycles were not considered in the present analysis. We obtained 269 pregnancies, divided as follows: 92, 77, 50 and 50 pregnancies at the first, second, third and fourth cycle, respectively. The multiple pregnancy rate was 16.7% (45 out of 269 pregnancies). In more detail, 36 of the multiple pregnancies were twins (twin pregnancy rate, 13.4%) and 9 were HOMP (HOMP rate, 3.3%). Seven out of nine HOMP occurred during the first six years of our reported data. In fact, in the last nine years we decided not to perform IUI in cases with detection of more than 3 follicles ≥16 mm on the day of hCG administration.

 

PRs for the entire sample are presented in Table 1. The CPR(i) after two cycles was 18.6% (95% CI, 0.166–0.217), and increased to 24.7% (95% CI, 0.227–0.283) and 31.4% (95% CI, 0.285–0.345) after cycles three and four, respectively; whereas the CPR(ii) was 18.3% (95% CI, 0.159–0.209), 23.7% (95% CI, 0.211–0.256) and 29.1% (95% CI, 0.263–0.321) after cycles two, three and four, respectively (Fig. 1). The difference in calculating the CPRs between the traditional and conservative estimate was 0.3% after the second cycle, 0.9% after the third, and 2.3% after the fourth cycle (Table 1). None of these differences was significant at the 5% level (χ2 = 0.032, df = 1; χ2 = 0.2319, df = 1; χ2 = 1.131, df = 1; respectively)

 

        

        

        .

 

Odds ratios for pregnancy in the second, third and fourth cycle compared to the first cycle was not significantly different from 1 at the 5% level of significance (χ2 = 0.057, df = 1; χ2 = 3.101, df = 1; χ2 = 0.385, df = 1; respectively).

 

The dropout rate was 3.4% (n = 31 couples) after the first cycle, 6.2% (n = 50) after the second cycle and 9.9% (n = 67) after the third cycle. The CDR after three cycles was 16% (148 out of 924 couples) of the overall sample. Considering pregnancies and dropouts, the couples who completed two, three and four cycles were 801, 674 and 557 respectively ( Table 1).

 

In order to verify our secondary objective, two hypothetical scenarios were computed, assuming dropouts similar to those reported in previous studies [5], [10] and [14]. Fig. 2 shows that our dropout rate per cycle, if compared to worse scenarios (CDR two times (scenario 2) or two and half times (scenario 3) greater than ours), permits recovery of 19 and 28 pregnancies, respectively.

 

        

 

Finally, female age at the beginning of treatment was categorized into three age groups: <35, n = 608; 35–39, n = 240; and >39, n = 76. The PR was significantly lower only in the oldest group. Considering the time of treatment onset at 5 year-intervals (1997–2001, n = 302, 2002–2006, n = 299, and 2007–2011, n = 323), the year of treatment was not significantly correlated to the outcome. Considering the causes of infertility, arranged in three categories (unexplained infertility, n = 408, male factor infertility, n = 354, and miscellaneous different causes, n = 88), the PRs were not significantly different ( Table 2).

 

        

 

4. Comment

A total of 924 couples who underwent 2956 cycles in a set of four IUI cycles over a long period of time (1997–2011) were analysed for the CPR and the CDR. After the third cycle, with a CDR of 16%, we found a CPR of 31.9% and 29.1% using the traditional and the conservative estimation, respectively. The CPR was in line with CPRs found in other similar studies [6] and [10]. The traditional estimate provides a very slight overestimation (2.3%) in CPR after four cycles and the difference between these two methodologies is quite small and not significant. This fact indicates the high reliability of the results, which is due to the very low number of patients who dropped out. We can affirm that our approach allows us to estimate the CPR more accurately, independent of the way it is calculated.

 

Considering our secondary objective, the hypothetical computation of the recovered pregnancies demonstrates that maintenance of a low CDR produces an increase in the CPR (Fig. 2).

 

Our CDR up to the third cycle was found to be lower than those reported in recent studies of IUI programs [5] and [14], and similar to that reported in countries where treatment is completely reimbursed [3] and [10]. Various studies have demonstrated that ceasing treatment is associated with financial concerns in countries where the treatment is not covered by a National Health System [15], [16] and [17], while a recent study, conducted in a reimbursed setting, affirmed that the decision to stop or continue treatment was not influenced by direct costs [3]. We can speculate that welfare policies providing for a reimbursement of expenditure reduce one of the possible factors affecting the dropout rate. However, in countries where this does not occur, pre-payment of a reasonable number of cycles, combined with the efforts of the physician to discuss the benefits of repeating IUI cycles, seems to be a feasible alternative.

 

In clinics where it is not possible to use a package of prepaid cycles, the reduction in dropout rates could be facilitated by means of accurate medical and psychological counselling in order to reduce the physical and psycho-social burden of treatment. It is crucial to explain that a couple should undergo a set of scheduled cycles to obtain high success rates; in obtaining this result, the cost of a cycle subsequent to the first could be reduced. The question may be raised of whether prepayment might raise ethical concerns, but we believe that the potential improvement of the CPR which can be obtained by the choice of the four-cycle program could overcome these concerns.

 

Furthermore, our results show a high grade of consistency among the different five year-intervals. We did not observe a significant reduction in clinical PR between the first and the other three cycles. This observation is comparable with some experiences [14] and contrasts with others [5] and [10]. Finally, in line with previous studies [14], the CPR was lower in patients older than 39 than in younger patients.

 

This study has two main limitations. It is an uncontrolled, retrospective, descriptive clinical study from a single clinic, with all the weaknesses associated with this approach. Secondly, in our study a cost-effectiveness analysis was not carried out, and the impact of this financial incentive should be better evaluated. In conclusion, the results of this study show that our “four-cycle program” produces a reduction in the CDR, allowing a better estimation of the CPR in IUI cycles.

 

Reduction of the CDR should represent an objective for all physicians who work in the field of reproduction, in order to give couples who are going to undergo IUI treatments more realistic chances of success and in order to increase the number of actual pregnancies. In line with these considerations, we believe that the CPR, instead of the PR per cycle, should become the reference point for communicating the outcomes to the couples, before IUI treatments.

 

Funding

This research was supported by funds provided by Centro Andros S.r.l., Palermo, Italy.