The relationship of total progressive motile sperm count with the outcome of IUI? An analysis of 5171 cycles

Abstract Background: The role of motile sperm count in intrauterine insemination (IUI) success rate is controversial. This retrospective cohort study performed among unselected infertile couples undergoing IUI was to explore the association between the total progressive motile sperm count (TPMSC) and the live birth rate (LBR) following IUI. Methods: The total cohort of 5363 cycles, 2666 infertile couples between January 2015 and December 2018 and finally 5171 cycles, 2647 couples were included for analysis in Sun Yat-sen memorial hospital of Sun Yat-sen University. The primary outcome was LBR per cycle. And the secondary outcome measure was clinical pregnancy rate (CPR) per cycle. Results: From the receiver operating characteristic (ROC) analysis of female age predicting live birth, female age cutoff was defined as 28 years. With a female age of ≤28 years, the CPRs were 11.5%, 14.9%, 16.1%, and 15.8% in quartile groups of pre-wash TPMSC, respectively. For the LBRs the values were 9.4%, 12.9%, 14.4%, and 11.3%, and there were also no significant differences in quartile groups of pre-wash TPMSC with ≤24 million (M), [24M–50M], [50M–97M], >97M. No statistically significant differences in the CPRs (p = .051) and LBRs (p = .088) were also observed in the quartiles groups of post-wash TPMSC. With a female age of >28 years, the CPR in couples with post-wash TPMSC ≤22.32 M was significantly lower than with post-wash TPMSC >81.0 M (p = .007). There was an obvious trend in which CPRs and LBRs increased with the post-wash TPMSC during the <81 M interval in women >28 years. Conclusions: The optimal female age cutoff for live birth was 28 years in IUI cycles. Pre-wash and post-wash TPMSC were not significantly associated with CPR and LBR per cycle. When female age >28 years, there was a better outcome with post-wash TPMSC >22.32 million.


Background
Intrauterine insemination (IUI) is a noninvasive and low-cost procedure to be considered in infertile couples, suffering from unexplained infertility, anovulation factor and mild male factor infertility, etc. Sperm wash is performed in order to get enough high-quality motile sperms for insemination. Theoretically, the total progressive motile sperm count (TPMSC) for insemination may affect the success rates of IUI. Recent studies [1][2][3][4][5][6] explored the predictive value of the total motile sperm count on clinical pregnancy and live births of IUI cycles., Bradley et al. [1]. from a retrospective study concluded that the average total motile sperm count in one ejaculate was an important factor, with a threshold value of 10 million in IUI. When the average total motile sperm count was below 10 million, IVF with ICSI was more cost-effective than IUI. Following this analysis, many clinics adopted the suggested threshold. In China, the majority of reproductive centers adhere to a cutoff level of TPMSC ≥10 million in IUI which is recorded in Human Assisted Reproductive Technical Specifications issued by the Ministry of Health of China. However, until now, there is no consensus on the criteria of a TPMSC cutoff for IUI [7][8][9][10][11][12].
In a recent retrospective analysis, including 655 cycles [13], it was concluded that pre-wash total motile sperm count (TMC) is a poor predictor of live birth. From the results, the female age influenced the cumulative live birth rate (LBR), and according to TMC, the LBR per cycle ranged from 5.1% to 12.5%. The optimal TPMSC was not presented.
In this study, we aimed to explore the relationship between the pre-wash and post-wash TPMSC for insemination for clinical pregnancy and LBRs in a cohort of infertile couples undergoing artificial insemination by husband (AIH)-IUI.

Semen analysis
After 3-7 days of abstinence, the semen was collected 1-2 h before insemination. Semen was examined according to the World Health Organization guidelines and processed using the density-gradient centrifugation method [14], The details are as follows (1): use two liquids, SpermGrade and SpermRinse, to prepare a gradient solution containing 90% SpermGrade and 45% SpermGrade in advance; (2) the gradient solution is equilibrated to room temperature; (3) using a sterile Pasteur pipette, add 1.0-2.0 ml of 90% gradient liquid to the sterile, sharp-bottomed centrifuge tube as the 'lower layer liquid' according to the volume of the semen, and gently add an equal volume of 45% gradient liquid above the liquid surface as the 'upper layer liquid'; (4) suction the completely liquefied semen with a sterile Pasteur pipette and slowly add it to a 15 ml conical centrifuge tube with gradient centrifugal fluid, and centrifuge at 500 g for 15 min; (5) discard the supernatant, add the precipitate to 2 ml G-IVF TM PLUS fertilization fluid (Vitrolife, Sweden), and centrifuge at 200 g for 4-10 min; and (6) aspirate the supernatant and leave 0.5 ml of sperm suspension for IUI use. The insemination was carried out by physicians, only. The TPMSC (million, M) was calculated by multiplying the total sperm concentration and volume (i.e. total volume for pre-wash, 0.5 ml in IUI for post-wash) with the progressive motility percentage determined pre-wash (before processing) and post-wash (after processing for insemination). Morphology was not used for calculations and analysis.

IUI treatment Ovarian stimulation protocol
If the patient had an irregular menstrual cycle, the ovarian stimulation cycle was preferred. Ovarian stimulation protocols included clomiphene citrate (CC, Codal Synto Ltd, Cyprus), CC in combination with gonadotropins, letrozole (LE, HENGRUI MEDICINE, China), letrozole in combination with gonadotropins, and gonadotropins only. Treatment with CC, letrozole or Gonadotropins alone, started on cycle day 5 for 5 days. Treatment with gonadotropins in combination with CC or letrozole started after completion of the oral agents and continued until desired follicular response. A urine luteinizing hormone (LH) test was performed daily by the patient once the dominant follicle had reached a size of ≥14 mm in diameter. Insemination was performed either on the day of a positive urine LH test or the day after trigger with HCG 5,000-10,000 U (urinary HCG, LIVZON, China or recombinant HCG, Merck, USA) in patients.

Natural cycle
If the patient had a regular menstrual cycle, the natural cycle was preferred. Ultrasound monitoring started on the 10-12th day of cycle and was combined with daily urine LH testing once the dominant follicle was ≥14 mm in diameter. Once the diameter of the dominant follicle reached 18 mm, the IUI was performed on the urine LH test positive day or the day after trigger with HCG 5,000-10,000 U in patients with urine LH negative testing.
In all IUI patients, the goal was to have one to three dominant follicles (diameter ≥ 14 mm).

IUI procedure
The volume of the final suspension of motile sperm was 0.5 ml and the insemination was carried out by a physician according to a standard protocol. A soft catheter was used for insemination, and the patient had bed rest for 30 min after the IUI procedure. 48 h later, the ovulation was confirmed by transvaginal ultrasound and followed by oral Dydrogesterone (Duphaston, Abbott Biologicals B.V., Netherlands) 10 mg, twice for 2 weeks until the urine pregnancy test was performed.

Outcome measures
Clinical pregnancy was defined by transvaginal ultrasonographic visualization of one or more gestational sacs when 4-5 weeks after insemination. This definition includes ectopic pregnancy. Live birth was defined as a viable delivery beyond 28 weeks' gestation. The primary outcome was LBR per cycle. And the secondary outcome measure was clinical pregnancy rate (CPR) per cycle.

Statistical analysis
Univariate analysis by Chi-square test was performed to identify factors that predict the chance of live birth. Univariate and multivariate logistic regression were used to assess the association between live birth and associated factors. Receiver operating characteristic (ROC) curves were used to explore the cutoff of age or TPMSC on the live birth. Patients were categorized into four groups according to TPMSC in quartiles. All the data were tested by one sample Kolmogorov-Smirnov test to confirm the normality. Non-normality distribution data were analyzed by non-parametric tests (Mann-Whitney U test and Kruskal-Wallis test). Categorical variables were assessed by chi-squared test or Fisher's exact test as appropriate. A p < .05 was considered statistically significant. IBM SPSS Statistics (version 22) was used to analyze data.

Results
A total of 5361 IUI cycles performed between January 2015 and December 2018 using homologous sperm-intrauterine insemination (IUI-H) were included in the analysis. A total of 133 cycles were excluded, including 69 cycles converted to IVF due to multi-follicular development (more than 3 dominant follicles); five couples (5 cycles) refused to undergo insemination owing to poor post-wash TPMSC; in a total of 51 cycles the male could not provide an ejaculate on the IUI day, and 8 cycles were canceled due to other factors such as fever. Moreover, a total of 57 cycles were excluded due to an incomplete data set.
Finally, a total of 5171 cycles in 2647 couples were included for analysis in this study.

Reproductive outcomes
During this study period, 945 couples (35.70%) had only one cycle of IUI. In 956 couples (36.12%) two cycles of IUI were performed. A total of 693 couples (26.18%) completed three cycles of IUI. There were 49 couples (1.85%) in whom 4 to 6 IUI cycles were performed, and 4 couples (0.15%) had 7 to 9 cycles of IUI. In total cycles, 570 clinical pregnancy cycles were obtained (CPR per IUI cycle 11.02%). Of them, 10 pregnant couples/cycles were lost to follow-up. There were 97 cycles ended in spontaneous abortion (miscarriage rate 17.01%) and 25 cycles recorded as ectopic pregnancies (ectopic pregnancy rate 4.38%). A total of 438 cycles were followed as live birth (LBR per IUI cycle: 8.47%; LBR per couple: 16.54%).
The univariate analysis showed that female age, female baseline FSH ≤ 10 U/L, female BMI ≤24 kg/m 2 , semen parameters such as pre-wash concentration, post-wash TPMSC, and concentration were significantly associated with live birth (Table 1).
Binary univariable and multivariable logistic regression indicated except female age, the mentioned variables above were not significantly associated with live birth since the OR values were very close to 1. (Suppl. Table 1). Female age was the one significantly associated with live birth. The cutoff female age for predicting live birth was defined as 28 years old by ROC analysis (Fig. 1). However, the AUC was 0.579 (95% CI 0.552-0.606, p = .000), and when female age was 28 years old, the sensitivity reached to 0.736 with specificity 0.386. From the characteristic comparison, two-thirds of female were beyond 28 years old. Ovarian induction protocol was dominant. Anovulation factor infertility accounted for 34.7% in the group with female age ≤28 years, which was higher than the group with female age >28 years ( Table 2).
When female age was beyond 28 years, according to the pre-wash TPMSC quartiles, the CPR and LBR were not statistically different in subgroups. However, in the subgroups of post-wash TPMSC quartiles, the CPR with post-wash TPMSC ≤22.32 million was statistically significantly lower than that with post-wash TPMSC >81.00 million (p = .007). And there was no significant difference in LBR 5.7%, 7.5%, 7.3% and 8.3% according to the post-wash TPMSC ≤22.32M, >22.32M and ≤46.5M, >46.5M and ≤81M, >81M (see Table 3).
The different values of post-wash TPMSC in the range of <81M for IUI were used to compare the LBR when the female age was >28 years. There was an obvious trend that CPR and LBR increased with post-wash TPMSC (Figure 2).

Discussion
In this large cohort analysis, we found that the optimal female age cutoff in live birth was 28 years in IUI cycles. Pre-wash and post-wash TPMSC were not significantly associated with CPR and LBR per cycle. When female age >28 years, there was a better outcome with post-wash TPMSC >22.32 million. There was an obvious trend that CPRs and LBRs increased with the post-wash TPMSC <81 million when female age >28 years. All these provide useful reference for IUI counseling.
IUI is commonly used for a part of infertile patients such as mild male factor infertility, female anovulation factor and unexplained infertility. Previous studies discussed several factors associated with IUI success rates; thus [3], concluded from a longitudinal cohort study (n = 1177 couples) that the pre-wash TMSC had a better correlation with the spontaneous ongoing pregnancy rate (SOPR) than other semen parameters. Van Voorhis et al. (2001) suggested that an average total motile sperm count of 10 million might be a useful threshold value for decision making whether a couple should be offered IUI or IVF. According to the ART specification in China, the post-wash TPMSC for IUI should not be less than 10 × 10 6 . So patients with low TPMSC (<10 million) on the IUI day will have to consent to continue the IUI or not. The TPMSC for insemination could be the second most important factor for IUI success except female age. Until now, few studies focused on the TPMSC. The value of post-wash TPMSC for predicting IUI outcomes is not well defined. Our analysis showed that both the pre-wash and the post-wash TPMSC are poor predictors of CPR and LBR in AIH-IUI. This finding corroborates the results of previous studies. Thus, Merviel et al. [2] reported a composite of positive prognostic factors in a retrospective study (n = 1038 cycles), concluding that TMC individually did not predict pregnancy. In another retrospective observational study, Lemmens et al. [15] performed an analysis on sperm morphology, TMC and TPMSC. The TPMSC had no predictive value. Moreover, a systematic review of 55 studies by Ombelet et al. [16] suggested that there is no optimal cutoff for IUI success as regards sperm parameters. Recently, Mankus et al. [13] performed a retrospective cohort study exploring the relationship between pre-wash total motile count (TMC) and LBRs in IUI cycles. The LBR per couple decreased to 7% in women over 37 years of age compared to 25% in women less than 37 years. No live births occurred with a TMC <2 million. Up to now, it is agreed that the female age is a key predictor of clinical pregnancy and live birth. The present retrospective cohort analysis revealed that the post-wash TPMSC significantly impacted the CPR and LBR when the age of the female was above 28 years.
The limitations of previous studies include small sample sizes and to the inability to identify clinically meaningful sperm count cutoff value. To clarify the relationship between post-wash TPMSC and CPR as well as LBR in IUI, we retrospectively analyzed a large single center sample. The sample size of the present study is seven times higher than that of prior studies, and moreover the present study specifically investigated the TPSMC. The ROC curve and logistic regression analysis did not show any correlation between pre-wash, post-wash TPMSC, clinical pregnancy and live birth per IUI cycle.
In the present analysis, when the post-wash TPSMC was less than 1 million, two live births were obtained in 94 cycles (LBR 2.13%). Furthermore, our study corroborates the preexisting publications concerning pre-wash and post-wash TPMSC being Figure 1. receiver operating characteristic curve for female age predicting live birth. note: aUc 0.579, 95% ci 0.552-0.606, p = .000, when female age was 28 years old, it can predict live birth with sensitivity 0.736 and specificity 0.386.  a poor predictor of clinical pregnancy and live birth. However, we observed that when female age >28 years, there was a better outcome with post-wash TPMSC >22.32 million.

Conclusions
Female age influenced the CPR. When the female age was >28 years, there was a trend that the CPR and LBR per cycle increased with the post-wash TPMSC. The findings of the present study could be used for counseling of the infertile couple as regards their chances of a live birth. For the low TPMSC couples, the female age should be added to the equation to decide when to convert to IVF/ICSI. To our knowledge, this is the largest study exploring the TPMSC and clinical pregnancy and LBRs. The limitation of this retrospective analysis is that it was performed in data from a single center, other factors influencing the IUI success rate such as the number of dominant follicles were not considered and the analysis might not be directly applicable to other settings and areas of the world.