Maternal underweight does not adversely affect the outcomes of IVF/ICSI and frozen embryo transfer cycles or early embryo development

Abstract Objective To compare assisted reproductive technology (ART) outcomes and preimplantation embryo development between underweight and normal-weight women. Methods This retrospective cohort study included 26 underweight women (body mass index [BMI] < 18.50 kg/m2) and 104 normal-weight women (BMI >20 and <24.9 kg/m2) who underwent a total of 204 in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) cycles and 358 fresh/frozen embryo transfers (ET) in our institution between January 2016 and December 2018. Statistical analyses compared selected ART outcomes (ovarian stimulation, fertilization, and pregnancy) between both weight groups. Morphokinetic and morphological parameters were also compared between 346 and 1467 embryos of underweight and normal-weight women, respectively. Results The mean ± standard deviation age of the underweight and normal-weight women was similar (31.6 ± 4.17 vs 32.4 ± 3.59 years; p = .323). There were no differences in the peak estradiol levels, the number of retrieved oocytes, the number of metaphase II oocytes, and the oocyte maturity rates between the two groups. The IVF/ICSI fertilization rates and the number of embryos suitable for transfer or cryopreservation were similar for both groups. All morphokinetic parameters that were evaluated by means of time-lapse imaging as well as the morphological characteristics were comparable between low and normal BMI categories. There were no significant differences in pregnancy achievement, clinical pregnancy, live births, and miscarriage rates between the suboptimal and optimal weight women. Conclusion Underweight status has no adverse impacts on the outcomes of IVF/ICSI with either fresh or frozen ET or on preimplantation embryo development and quality.


Introduction
The association between overweight (BMI >25 kg/m 2 ) and impaired fertility [1], poor assisted reproductive technology (ART) outcomes [2], and increased rates of miscarriage and obstetric risks in spontaneous [3][4] and ART pregnancies [5][6] is well-recognized. Similarly, underweight women (BMI <18.50 kg/m 2 ) have higher rates of infertility compared to women with a normal BMI (18.50-24.99 kg/m 2 ) [7]. A low BMI is also associated with an increased risk of miscarriage [8] and obstetric complications [9] in spontaneous pregnancies. However, the association between prepregnancy subnormal body weight and ART outcomes has not been established [10][11].
Chronic energy and nutrient deficiencies can disrupt the hypothalamic-pituitary-gonadal (HPG) axis and lead to irregular menstrual cycles, ovulatory dysfunction, and thereby reduced reproductive function in underweight women [7]. However, the mechanisms by which underweight might affect ART outcomes, a process that bypasses the HPG axis, are far less clear and are probably related to multiple endocrine, adipokine, and metabolic alternations that can affect follicle growth, oocyte maturation, early embryo development, and endometrial receptivity.
Previous studies explored the effect of female underweight on ART outcomes included a variety of pathophysiologies underlying infertility, and most of those studies focused upon fresh embryo transfer (ET). Although several studies found impaired ART outcomes, known to be associated with impaired embryo development and low embryo quality, among underweight women undergoing ART [11][12][13], only one study explored embryo morphokinetics in this context [13].
Given the lack of consensus over the effect of maternal underweight on ART outcomes, the aim of the current study was to investigate the association between low BMI and ART outcomes in women undergoing fresh and frozen cycles for IVF/ intracytoplasmic sperm injection (ICSI). To strengthen the data, early embryo development and quality were also examined.

Study population and participant recruitment
This retrospective, single center research was undertaken in a university affiliated tertiary hospital between January 2016 and December 2018 and was approved by the local Institutional Review Board (#0977-20-TLV). Morphological and morphokinetic developmental patterns of 1813 embryos, derived from 130 women, were assessed, and pregnancy outcomes from 358 ET cycles that included 513 transferred embryos were determined.
To prevent any influence of infertility factors on the outcome of IVF, we limited the analysis to the data of women who underwent IVF due to genetic indications for the performance of preimplantation genetic diagnosis (PGD: 56 cases, 43%), unexplained infertility (61 cases, 47%), and mechanical factor infertility (13 cases, 10%). We excluded women who presented with the following: uterine malformation, endometriosis, or hydrosalpinx; a history of uterine and ovary surgery; any thyroid disease, diabetes, or autoimmune disease; more than three previous IVF/ICSI treatments; a baseline follicle-stimulating hormone (FSH) level >12 IU/L; being a Fragile X carrier; and, infertility due to male factor.
The BMI was calculated for all participants and they were divided into underweight (BMI <18.5) and normal-weight (BMI >20 and <24.9) groups. Women with a BMI >18.5 and <20 were not included in this study in order to clearly differentiate between 'underweight' and 'normal'-weight.

Data collection
All relevant data were collected from the hospital computerized database. A serum beta human chorionic gonadotropin (b-hCG) level >25 IU/I was taken as positive for pregnancy. A clinical pregnancy was confirmed by the observation of an embryo pulse on transvaginal ultrasound scanning at 6-12 weeks of gestation. Early miscarriage was diagnosed when a previously positive pregnancy test become negative before ultrasonographic detection of an embryonic pulse in the sixth week of pregnancy or later. Miscarriage was defined as a loss of clinical pregnancy before 12 full weeks of gestation. Live birth was defined as a live neonate born after 24 weeks of gestation. Pregnancy, clinical pregnancy, and live birth rates were on a 'per ET' basis. Early miscarriage and miscarriage rates were calculated per pregnancy.
Ovarian stimulation, fertilization, embryo culture and embryo transfer Optional protocols and laboratory methods were elaborated previously by our group [14]. All embryos were incubated in the integrated EmbryoScope TM time-lapse monitoring system (EmbryoScope TM ; UnisenseFertiliTech A/S, Aarhus, Denmark, Vitrolyfe) from the time of fertilization until ET or freezing. Either ET or cryopreservation was carried out 2-6 days following oocyte retrieval. Endometrial preparation in the cases of frozen ET (FET) was performed with modified natural or hormonally substituted cycle protocols [15]. Luteal support with progestin supplement was continued in all cases until there was a negative b-hCG result or the ninth week of pregnancy. Serum b-hCG levels were confirmed on day 14 after ET.

Time-lapse monitoring of embryo morphokinetics and morphology assessment
Embryo scoring and selection with time-lapse monitoring were performed by analysis of the time-lapse images of each embryo with software developed specifically for image analysis (EmbryoViewer workstation; UnisenseFertilitech A/S). Embryo morphology and developmental events were recorded to demonstrate the precise timing of the observed cell divisions in correlation to the timing of fertilization: time of pronuclei fading (tPNf), cleavage to a 2-blastomere (t2), a 3-blastomere (t3), a 4blastomere (t4), etc. until an 8-blastomere (t8) embryo. The other analyzed parameters were the lengths of the second and the third cell cycles (cc2 and cc3), and the synchrony in the division from 3 to 4 cell (s2) and 5 to 8 cells (s3). Scores were allocated to day 3 embryos by means of the KIDScore algorithm [16]. Conventional morphology of the embryos was studied on day 3, considering the number of blastomeres, the symmetry among blastomeres, and the degree of fragmentation. The embryos were scored from grade 1 (high quality) to grade 4 (poor quality) accordingly [17].

Statistical analysis
Data were analyzed with SPSS, version 25.0 (SPSS, Inc., Chicago, IL, USA). The clinical characteristics as well as the ART and pregnancy data and outcomes were summarized as mean þ standard deviation (SD), or number of responders (percentage) according to the variables. Significance was tested with the t-test, Mann-Whitney U test, v2, and Fisher's exact test as appropriate. The effect of BMI status on morphokinetic parameters was assessed by a mixed model's analysis. A Generalized Linear Mixed Model (GLMM) analysis was performed to assess the impact of the confounding of ET strategies, including ET protocol type, number of embryos transferred and day of ET, on live birth rate. A p value of < .05 was considered significant.

Clinical characteristics of the study participants
In total, 130 women underwent 204 conventional IVF, ICSI, or IVF/ICSI cycles and were included in this study. Underweight (BMI <18.5) women comprised 20% (26 out of 130) of our study participants. Female age was similar for underweight and  (Table 1).

Ovarian stimulation data and outcomes
The 26 underweight women and 104 normal-weight women underwent 42 and 162 OPU cycles, respectively. The ovarian simulation data and outcomes of both groups are summarized in

Morphokinetic and morphological characteristics
Three-hundred forty-six embryos of underweight women were compared morphokinetically to 1467 embryos of normal-weight women. The mean timings of tPNf, and t2 to t8, along with cc2, cc3, and s2 to s3 were not significantly different in the maternal BMI group comparison ( Table 3).
The KIDScore was calculated for 278 embryos from underweight women and 1305 embryos from normal-weight women. All of those embryos were available for embryoscopic analysis at 66 h and were fit to be graded according to the model in which each embryo receives a score between 1 and 5 (1 indicating the lowest potential for pregnancy and 5 indicating the highest). The mean scores were similar for embryos from the underweight and from the normal-weight women (3.7 ± 1.50 vs. 3.63 ± 1.51, respectively; p ¼ .670) ( Table 3). There was also no significant group difference between the proportion of embryos graded either 4 or 5 or 2 or less (p ¼ .593) ( Table 4).
Additionally, we performed a conventional morphological evaluation on day 3 and each embryo received a score between 1 and 4 (1 indicating the highest potential for pregnancy and 4 indicating the lowest). The Morphology score was calculated for 283 embryos from underweight women and 1309 embryos from normal-weight women. The mean scores were similar for embryos from the underweight and the normal-weight women (1.82 ± 1.04 vs. 1.95 ± 1.10, respectively; p ¼ .230) ( Table 3). There was also no significant group difference between the proportion of embryos graded 1 or 2, or those graded 3 or more (p ¼ 0264) (Table 4). Finally, no significant group difference was found in the percentage of embryos whose development was stopped before day 3 (8.4% vs. 8.2%; p ¼ .910) ( Table 3).

Pregnancy outcomes
One hundred and one embryos from underweight women and 412 embryos from normal-weight women were transferred, and the pregnancy outcomes were assessed (

Discussion
There is no consensus over the effect of maternal underweight on IVF outcomes. The value of this information is its contribution in establishing recommendations, if any, regarding the lower weight threshold for optimal IVF outcomes. The present study has demonstrated that underweight women undergoing IVF treatment do not appear to be at greater risk of impaired IVF outcomes than normal-weight women. Two meta-analyses summarized the association between maternal underweight and pregnancy outcomes. The first one [8] concluded that maternal subnormal body weight prepregnancy is associated with a slightly increased risk of clinical miscarriage compared with normal-weight women. However, this meta-analysis included spontaneous and ART-induced pregnancies, and the types of pregnancy among underweight women did not yield significantly different outcomes. Furthermore, the conclusions were limited by not showing any association between prepregnancy subnormal body weight and other critical outcomes of ART. The second meta-analysis [18] focused solely upon IVF pregnancies, exploring the association between subnormal body weight prepregnancy and multiple obstetrical outcomes. The authors concluded that there were no differences in miscarriage or live birth rates between underweight and normal-weight patients, but clinical pregnancy rates were slightly impaired among the underweight women.
However, none of the 38 studies, which were included in this meta-analysis, independently demonstrated a significant reduction in clinical pregnancy.
Inconsistency remained after Xiong et al.'s [18] meta-analysis as well. Like our results, however, most of these recent studies failed to find any association between underweight and negative IVF outcomes [10,[19][20][21][22]. Contrarily, Tang et al. found that being underweight was linked to reduced implantation, clinical pregnancy, and ongoing pregnancy rates in women undergoing FET-based IVF [11]. We consider that the influence of biological differences among different ethnicities might explain these differences in outcomes between studies. Indeed, Tang et al.'s study of a Chinese population observed a negative influence of low BMIs among Asian women on IVF outcomes, and Cai et al. [12] had earlier reported that a low BMI was associated with reduced live birth rates and increased miscarriage rates compared with normal-weight in a Chinese population. The other studies [10,[19][20][21][22], as well as ours, had been conducted in a non-Asian population, and none observed any correlation between low prepregnancy maternal weight with IVF outcomes. An exception is the study of Bartolacci et al. [13], who reported a higher miscarriage rate in underweight compared with normal-weight women undergoing IVF and suggested that the mechanisms underlying a higher risk of miscarriage included differences in embryo quality and endometrial receptivity. Those authors [13] were the first to examine the effect of underweight on embryo morphokinetics and found no significant differences in time-lapse imaging parameters between underweight and normal-weight women. Similar findings were recently demonstrated in Bellver et al.'s paper [23]. This study compared 965 embryos derived from 140 underweight women with 12,446 embryos derived from 1989 normal- Values are presented as number (%). A p value of < .05 was considered significant. KID and morphological scores were allocated to day 3 embryos. BMI: body mass index. Values are presented as % (standard deviation). A p value of < .05 was considered significant. Pregnancy, clinical pregnancy, and live birth rates were on a "per embryo transfer" basis.
Early miscarriage and miscarriage rates were calculated per pregnancy. Ã A Generalized Linear Mixed Model (GLMM) analysis was performed to assess the impact of the confounding of ET strategies, including ET protocol type, number of embryos transferred and day of ET, on live birth rate.
weight women and found no significant differences in all examined morphokinetic parameters between these two groups. Contrarily, Kassi et al. [24] compared 56 embryos derived from 17 underweight women with 1252 embryos derived from 327 normal-weight women and observed that embryos from underweight women have a faster time to the 8-cell stage than normal weight. However, the possible reasons for these findings and the potential implications of faster preimplantation embryo development and pregnancy/obstetric outcomes were not discussed.
Our results confirmed that there is no significant relationship between female underweight and embryo developmental kinetics, thereby supporting the similar clinical pregnancy and miscarriage rates we found between the two weight groups. The discrepancy between Bartolacci's study and ours regarding miscarriage rate could be due to different sample sizes and/or methodologies. The women in Bartolacci's study ($36 years) were older than the participants in the current study ($32 years). Advanced maternal age may have undesirable effects upon endometrial receptivity and thereby reduce embryo implantation and successful pregnancy rates [25]. Bartolacci et al., unlike us, had included various infertility causes that are known to impair IVF outcomes. The differences between the results of our study and theirs may also be due to a synergistic effect of these factors and underweight. However, our results should be taken with caution due to the small subgroups. In the case of early miscarriage rate if we had 197 underweight women and 788 normal-weight women the results would be significantly different and agree with what is found in previous studies [8,13].
Our study has several important limitations. (1) it is retrospective in design, which reduces its direct application to clinical practice. (2) the statistical analyses were performed in relatively small subgroups, which could compromise the strength of the conclusions, thus calling for further studies on larger populations. (3) we did not examine the effects of prepregnancy maternal underweight on obstetric outcomes following IVF, and this, too, awaits further studies.
In conclusion, we demonstrated that maternal underweight is not associated with poorer IVF/ICSI/FET outcomes or with impairment of preimplantation embryo development, thus supporting the findings of others in this ongoing controversy. However, a suboptimal BMI before IVF was found to be an independent risk factor for many adverse obstetric outcomes, including preterm birth, intrauterine growth restriction, and small for gestational age neonates [19,26,27]. Therefore, dietary and lifestyle counseling intended to achieve optimal body weight prior to IVF is recommended in order to improve not only IVF pregnancy rates but also the chances of an uncomplicated pregnancy, childbirth, and postnatal period.

Author contributions
D.H. and H.A. were involved in the conception, study design, data acquisition and analysis and, manuscript preparation. Y.K. collected and managed the database and provided her expertise in the critical reading of the manuscript. N.S., and E.H.H. were involved in the conception, managed the database and, provided their expertise in the critical reading of the manuscript. S.L. managed and statistically analyzed the database. F.A. contributed to the interpretation of the data and provided their expertise in the critical reading of the manuscript. All contributors reviewed and edited the manuscript and gave their approval of the final version.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.