Human Reproduction, Vol.31, No.2 pp. 385– 392, 2016 Advanced Access publication on December 17, 2015 doi:10.1093/humrep/dev298
ORIGINAL ARTICLE Infertility
Reproductive outcomes in oocyte donation cycles are associated with donor BMI E.R. Cardozo 1,2,3,*, A.E. Karmon 1,2,3, J. Gold 1,3, J.C. Petrozza 1,2,3, and A.K. Styer1,2,3 1 Vincent Reproductive Medicine and IVF, Vincent Department of Obstetrics and Gynecology, Massachusetts General Hospital, Boston, MA 02114, USA 2Department of Obstetrics, Gynecology, and Reproductive Biology, Harvard Medical School, Boston, MA 02115, USA 3 Present address: Vincent Reproductive Medicine and IVF, Vincent Department of Obstetrics and Gynecology, Massachusetts General Hospital, Yawkey 10A, 55 Fruit Street, Boston, MA 02114, USA
*Correspondence address. E-mail:
[email protected]
Submitted on June 4, 2015; resubmitted on September 24, 2015; accepted on November 6, 2015
study question: When adjusting for recipient BMI, is donor body mass index (BMI) associated with IVF outcomes in donor oocyte IVF cycles?
summary answer: Increasing oocyte donor BMI is associated with a reduction in clinical pregnancy and live birth rates. what is known already: Increased BMI has been associated with suboptimal reproductive outcomes, particularly in assisted reproductive technology (ART) cycles. However, it remains unclear if this association implies an effect of BMI on oocyte quality and/or endometrial receptivity.
study design, size, duration: A retrospective cohort study of two hundred and thirty five consecutive fresh donor oocyte IVF cycles from 1 January 2007 through 31 December 2013 at the Massachusetts General Hospital (MGH) Fertility Center. participants/materials, setting, methods: Analyses included a total of 202 oocyte donors and 235 total cycles. Following adjustments for recipient BMI, the relationship between donor BMI (categorized into quartiles) and IVF outcomes was assessed. main results and the role of chance: In the entire (anonymous and known) donor population, a reduced odds of clinical pregnancy (P-trend ¼ 0.046) and live birth (P-trend ¼ 0.06) was observed with increasing BMI quartile. Compared with quartile 1 (BMI 17.8 –21.1), odds ratio (OR) (95% CI) of clinical pregnancy was 0.9 (0.4 –2.0), 0.5 (0.2 –1.1) and 0.5 (0.2 –1.1), and OR of live birth was 1.1 (0.5 –2.6), 0.6 (0.3– 1.2) and 0.6 (0.3 –1.2) for quartiles 2 through 4 respectively. In anonymous donors only, the odds of clinical pregnancy (P-trend ¼ 0.02) and live birth (P-trend ¼ 0.03) also declined as BMI quartile increased. Compared with quartile 1 (BMI 17.8 –21.1), odds ratio (OR) (95% CI) of clinical pregnancy was 0.7 (0.3 –1.7), 0.5 (0.2–1.1) and 0.4 (0.1 –0.9), and OR of live birth was 0.9 (0.4 –2.2), 0.5 (0.3 –1.2) and 0.4 (0.2 –1.1) for quartiles 2 through 4 respectively. limitations, reasons for caution: Limitations include the retrospective design, sample size and data from a single institution. Clinical application may not be limited to oocyte donors, though caution should be used prior to applying these principles to the general population. Data should not be interpreted to mean that all oocyte donors should be restricted to a BMI of less than 21.2 kg/m2.
wider implications of the findings: Following adjustments for the respective BMI of the oocyte donor and recipient, this study demonstrates an association of preconception BMI with subsequent IVF outcomes. The observations of this study are consistent with prior animal studies, suggest a possible effect of BMI at the oocyte level prior to fertilization and implantation, and warrant further investigation. study funding/competing interests: None. Key words: body mass index / obesity / oocyte / oocyte donor / in vitro fertilization / oocyte recipient / reproductive outcomes / oocyte quality
& The Author 2015. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email:
[email protected]
386
Introduction Excess weight has a particularly deleterious impact in women, resulting in cardiometabolic health risks as well as adverse reproductive health outcomes (Practice Committee of American Society for Reproductive Medicine, 2008), of which most women are unaware (Cardozo et al., 2012, 2013). In addition to being a risk factor for infertility (Rich-Edwards et al., 1994), women who are overweight and obese undergoing assisted reproductive technologies (ART) require higher doses of gonadotrophins (Maheshwari et al., 2007), have fewer oocytes retrieved (Zhang et al., 2010), and have lower rates of fertilization (Zhang et al., 2010), implantation (Moragianni et al., 2012), and pregnancy (Wang et al., 2000; Loveland et al., 2001; Maheshwari et al., 2007; Rittenberg et al., 2011). While there are substantial data indicating that excess weight has a negative impact on the reproductive outcomes of ART cycles, it remains unclear if this effect is exerted at the level of the oocyte or the endometrium (Cardozo et al., 2011). Classically, the oocyte donation model has been used to differentiate the effect of obesity on oocyte/embryo quality and endometrial receptivity; however, conflicting results remain. Several studies suggest that unfavorable ART outcomes in obese women are caused by suboptimal oocyte quality (Cano et al., 1997; Wattanakumtornkul et al., 2003; Zenke and Chetkowski, 2004; Styne-Gross et al., 2005; Luke et al., 2011). In contrast, many investigators highlight the negative impact of obesity on endometrial receptivity (Bellver et al., 2003, 2007, 2013; DeUgarte et al., 2010). These prior studies have assumed that donors have no comorbidities and have a normal body mass index (BMI), and have thus excluded the impact of the preconception environment by only accounting for the recipient BMI and not adjusting for donor BMI in their analyses. Murine models demonstrate that obesity negatively impacts the quality and developmental competence of oocytes and embryos, and this effect of BMI persists even when the oocyte or embryo is removed from the obese host environment (Minge et al., 2008; Luzzo et al., 2012). Therefore, investigation of the potential impact of donor obesity on IVF outcomes in donor-recipient cohorts will provide invaluable insight into the possible effect of BMI prior to conception in humans. The objective of this study was to assess the association of BMI on donor oocyte in vitro fertilization (IVF) cycle outcomes by utilizing an oocyte donation model that accounts for the BMI of donor and recipient respectively.
Materials and Methods Study population Two hundred and forty-seven consecutive donor oocyte IVF cycles from 1 January 2007 through 31 December 2013 at Massachusetts General Hospital (MGH) Fertility Center were reviewed. All fresh donor oocyte cycles in which controlled ovarian hyperstimulation was initiated and had donor BMI data available, including subsequent fresh oocyte donation cycles by the same donor, were included. Of the 247 consecutive fresh donor oocyte cycles reviewed, analyses included a total of 202 oocyte donors and 235 total cycles. Anonymous donors (unknown to the recipient) are selected by recipients via a third party agency and known donors (typically a relative or friend of the recipient) are specifically selected by the recipient. All donors undergo a multidisciplinary medical and psychological evaluation to determine eligibility for donation as described previously (Practice Committee of American
Cardozo et al.
Society for Reproductive Medicine and Practice Committee of Society for Assisted Reproductive Technology, 2013).
Stimulation protocols Oocyte donors Oocyte donors underwent one of three stimulation protocols as previously described by our center (Styer et al., 2008, 2015; Cardozo et al., 2015): the long luteal phase GnRH agonist (GnRH-a) protocol (‘Low Dose Luteal Lupron’ or LDLL), the follicular phase GnRH agonist (Flare) protocol, or the GnRH antagonist protocol.
Recipients The programmed hormone replacement regimen consisted of oral contraceptive pill (OCP) pretreatment followed by pituitary down-regulation with daily LupronTM (leuprolide acetate, 0.5 mg/d SC; TAP Pharmaceuticals Inc., Lake Forest, IL, USA) starting the last 5 days of OCP use, followed by transdermal estradiol (Vivelle-DotTM ; Novartis Pharmaceuticals, East Hanover, NJ, USA) for 2– 3 weeks until the day of donor hCG trigger. On the day of donor oocyte retrieval, the recipient initiated i.m. micronized progesterone (50 mg daily; Progesterone in Sesame Oil; Watson Pharmaceuticals Inc., Parsippany, NJ, USA), vaginal micronized progesterone (100 mg per vagina twice daily; Endometrin; Ferring Pharmaceuticals, Tarrytown, NY, USA), and transdermal estradiol (0.2 mg every other day; Vivelle-Dot; Novartis Pharmaceuticals, East Hanover, NJ, USA) which was continued until 13 weeks gestation.
Data collection and statistical analysis Data were collected through review of electronic medical record for all oocyte donors and donor oocyte recipients. Primary study outcomes were implantation (number of gestational sacs visualized per number of embryos transferred), clinical pregnancy (the presence of at least 1 gestational sac visualized at 6 weeks estimated gestational age), and live birth (birth of a viable infant on or after 24 weeks estimated gestational age) per cycle start. Secondary study outcomes included donor peak estradiol level, total number of oocytes retrieved from oocyte donor, number of mature oocytes retrieved, fertilization rate (defined as number of mature oocytes inseminated divided by number of two pronuclear embryos seen on the day following in vitro fertilization), and cleavage rate (percent of embryos demonstrating ≥2 blastomeres 2 days following in vitro fertilization). Statistical analysis was performed using SAS Statistical software version 9.3 (SAS Institute, Cary, NC, USA). BMI of all donors was divided into quartiles. All data are presented by donor BMI quartile in order to understand the nature of the relationship between BMI and outcomes without assuming it is linear, as well as to detect odds ratios per group and not just per incremental BMI unit. Generalized linear mixed models were used to assess cycle characteristics and all results were adjusted for donor age, except for recipient lining and pattern, which were adjusted for recipient age. Results are presented as adjusted mean (95% confidence interval) or adjusted probability (95% confidence interval). Cycle outcomes were analyzed with logistic regression using generalized estimating equations (GEE) to take advantage of all available cycles while taking into account within-person correlations in cycle outcomes (controlling for repeated cycles) (Fitzmaurice et al., 2004) and presented as an adjusted odds ratio, with the exception of implantation rate which was analyzed using generalized linear mixed models and presented as an adjusted rate. Outcomes were adjusted for donor age, recipient age, recipient BMI, male factor, number of prior cycles in donor, donor antral follicle count, and donor status (known versus anonymous). The variables which were adjusted for were determined based on those that had P-values of less than or equal to 0.2 in univariate analysis (Table I), which included recipient BMI, male factor infertility, donor’s number of prior cycles, and donor status (known versus anonymous), as well as those
387
Oocyte donor BMI and reproductive outcomes
Table I Characteristics of donor and recipient at donor’s first cycle at MGH, by donor BMI quartile, mean (standard deviation) or percentage (frequency). BMI Q1 17.8–21.1
BMI Q2 21.2– 22.8
BMI Q3 22.9– 25.2
BMI Q4 25.3–34.0
P-value
............................................................................................................................................................................................. Women (n)
46 2
51
54
51
Donor BMI (kg/m )
19.7 (0.9)
21.9 (0.5)
23.9 (0.7)
27.7 (2.0)
,0.01
Donor age (years)
26.2 (3.3)
27.7 (4.1)
26.6 (3.0)
27.4 (4.3)
0.36
Donor AFC
18.3 (5.7)
19.1 (6.4)
20.4 (7.9)
18.9 (7.7)
0.34
6.0 (1.5)
6.0 (1.6)
6.3 (1.6)
6.1 (1.7)
0.71
Donor D3 FSH (IU/l) Total number of previous cycles*
1.5 (1.4)
1.4 (1.4) 19.6% (10)
1.3 (1.5) 13.0% (7)
1.0 (1.2)
0.02
23.5% (12)
0.11
Known donor, % (n)
6.5% (3)
Recipient age (years)
41.4 (4.9)
40.7 (4.7)
41.1 (4.6)
41.8 (4.7)
0.56
Recipient BMI (kg/m2)
22.6 (2.8)
23.7 (3.8)
24.6 (4.7)
25.7 (5.0)
,0.01
Uterine factor, % (n)
2.2% (1)
2.0% (1)
0.0% (0)
3.9% (2)
0.55
Male factor, % (n)
4.4% (2)
13.7% (7)
20.4% (11)
19.6% (10)
0.10
BMI, body mass index; AFC, antral follicle count; FSH, follicle stimulating hormone. *Based on donor’s last cycle at MGH.
which are known to have an impact on IVF cycle outcomes based on previously published literature, including donor age (Barton et al., 2010), recipient age (Yeh et al., 2014), and donor antral follicle count (Vrontikis et al., 2010). A P-trend of ,0.05 was considered significant.
Ethical approval This study was approved by the Partners Human Research Committee (Institutional Review Board of Partners HealthCare).
Results A total of 202 oocyte donors underwent 235 total cycles. Exclusion criteria included cancelation prior to cycle start, missing BMI data, and conversion to cryopreservation with no fresh embryo transfer. At the time of first donor oocyte cycle at MGH, the mean (range, SD) donor age (years) was 27.0 (21.6–42.3, 3.7) and the mean recipient age was 41.2 (29.9– 49.7, 4.7). Mean (range, SD) donor BMI was 23.4 kg/m2 (17.8–34.0, 3.1) and mean (range, SD) recipient BMI was 24.2 kg/m2 (17.8–40.1, 4.3). The mean (range, SD) number of prior fresh cycles per donor at the time of their last cycle was 1.4 (0.0 –5.0, SD 1.4). Of the 235 oocyte donation cycles, 230 (97.9%) used a low dose luteal downregulation protocol, 3 (1.3%) used a GnRH antagonist down-regulation, and 2 (0.9%) used a GnRH flare. Total clinical pregnancy rate per cycle start (all cycles) was 61.7%, and overall live birth rate per cycle start was 54.9%. The BMI of all donors was categorized into quartiles with approximately equal numbers of donors per quartile (at the boundaries there were several donors with the same BMI who could not be assigned to different groups), while recipient BMI remained a continuous variable. Subsequent analyses were subdivided into the following donor BMI quartiles: Q1 (17.8–21.1 kg/m2), Q2 (21.2–22.8 kg/m2), Q3 (22.9– 25.2 kg/m2) and Q4 (25.3–34.0 kg/m2). Characteristics of donors and recipients at the time of donor’s first cycle, categorized by donor BMI quartile, are presented in Table I. Total number of previous cycles decreased with increasing BMI quartile (P ¼ 0.02). Except for BMI,
there were no differences in donor or recipient characteristics between quartiles. There was a positive correlation between donor BMI and recipient BMI for all cycles (R ¼ 0.31, P , 0.01) and for anonymous cycles only (R ¼ 0.26, P , 0.01). When each of the cycle characteristics listed in Table I were compared between known and anonymous donors, known donors were more likely to be older (P , 0.01), have a higher BMI (P , 0.01), have a higher day 3 FSH (P ¼ 0.02), and completed fewer prior cycles (P . 0.01). All other cycle characteristics were similar between groups (data not shown). The odds ratios for clinical pregnancy and live birth of our covariates, which are included in the main model, are presented as Supplementary Table SI. Additional cycle characteristics are presented in Table II. All results are adjusted for donor age, with the exception of recipient endometrial thickness and pattern, which are adjusted for recipient age, and use of ICSI, which is not adjusted for age. No difference was observed between quartiles with respect to total gonadotrophin dose used, day of hCG administration, thickness or pattern of recipient lining on day of trigger injection, use of ICSI for insemination, fertilization rate, cleavage rate, number of embryos transferred, or likelihood of a Day 5 embryo transfer. A decrease in peak serum estradiol level was observed as BMI quartile increased (P-trend ¼ 0.07), and mean number of eggs retrieved decreased with increasing BMI quartile (P-trend ¼ 0.06) but neither reached statistical significance. Nine cycles were canceled after stimulation start due to poor response. The median BMI for canceled cycles was 28.2 kg/m2 versus 22.9 kg/m2 for non-canceled cycles. Seventeen spontaneous abortions occurred in 235 cycles for a spontaneous abortion rate of 7.2%. There was no difference in median BMI for those who had a spontaneous abortion versus those who did not (22.9 kg/m2, interquartile range 4.1 versus 23.5 kg/m2, interquartile range 3.9; P ¼ 0.81). Table III presents cycle outcomes by donor BMI quartile (Quartile 1 is the reference). Following adjustments for donor age, recipient age, recipient BMI, male factor infertility, donor’s number of prior IVF cycles, donor antral follicle count, and donor status (known versus anonymous), no difference was observed in implantation rate or
388
Cardozo et al.
Table II Cycle characteristics by donor BMI quartile, adjusted* mean or adjusted* probability (95% CI). BMI Q1 17.8–21.1
BMI Q2 21.2– 22.8
BMI Q3 22.9–25.2
BMI Q4 25.3–34.0
P-trend
............................................................................................................................................................................................. Number of cycles
55
60
65
55
Total gonadotrophin dose (IU), mean
1845 (1625–2065)
1848 (1635– 2061)
1900 (1709– 2090)
2054 (1807–2300)
0.18
Peak E2 (pg/ml), mean
2680 (2349–3010)
2738 (2465–3010)
2482 (2186–2778)
2366 (2068–2663)
0.07
Day of donor hCG, mean
12.1 (11.2–13.1)
11.9 (11.0 –12.8)
12.1 (11.3– 13.1)
12.2 (11.3–13.2)
0.73
Recipient lining thickness on day of donor trigger (mm), mean
10.2 (9.6–10.8)
9.8 (9.1– 10.5)
9.9 (9.3–10.4)
10.0 (9.2–10.7)
0.68
Recipient lining pattern, multilayered, %
94.2 (82.3–98.2)
96.3 (85.5 –99.1)
96.6 (86.4– 99.2)
92.9 (81.2–97.5)
0.70
Number of eggs retrieved, mean
15.1 (13.3–17.2)
14.8 (13.1 –16.7)
13.8 (12.2– 15.6)
12.9 (11.3–14.7)
0.06
ICSI, %
41.9 (29.1–55.9)
36.7 (25.1 –50.2)
40.1 (28.4– 53.0)
45.4 (32.3–59.3)
0.59
Fertilization rate (normal, 2PN), %
73.4 (68.1–78.2)
72.7 (67.6 –77.1)
74.2 (69.2– 78.7)
71.5 (65.7–76.6)
0.69
Cleavage rate, %
98.7 (96.7–99.5)
99.5 (98.2 –99.9)
99.2 (97.7– 99.7)
98.9 (96.9–99.6)
0.96
1.7 (1.4–2.1)
1.8 (1.4– 2.1)
1.6 (1.3–2.0)
1.6 (1.3–2.0)
0.41
62.9 (47.5–76.0)
53.4 (39.3 –66.9)
58.0 (43.6– 71.2)
62.2 (46.2–75.9)
0.88
Number of embryos transferred, mean Day 5 transfer (versus Day 2 or 3), %
E2, estradiol; hCG, human chorionic gonadotrophin; ICSI, intracytoplasmic sperm injection; 2PN, two pronuclear embryo observed on the day following in vitro fertilization (i.e. normal fertilization). *Results are adjusted for age (donor or recipient as applicable). ICSI rate is not adjusted.
Table III Cycle outcome by donor BMI, adjusted* OR (95% CI) or adjusted* rate (95% CI). BMI Q1
BMI Q2
................................ ................................... Crude
Multi variate
Crude
Multi variate
BMI Q3
...................................
Crude
Multi variate
BMI Q4
...................................
Crude
Multi variate
P-trend Multi variate
............................................................................................................................................................................................... Cycles (number)
55
55
60
60
65
65
55
55
–
Biochemical pregnancy
41 (74.5%)
Reference
46 (76.7%)
1.0 (0.4–2.5)
44 (67.7%)
0.7 (0.3–1.5)
32 (58.2%)
0.4 (0.2– 1.0)
0.03
Implantation rate
53.1%
52.9% (38.8– 66.6)
55.0%
52.3% (39.6– 64.8)
49.2%
47.6% (34.7–60.9)
43.2%
42.3% (29.1 –56.8)
0.24
Clinical pregnancy
38 (69.0%)
Reference
41 (68.3%)
0.9 (0.4–2.0)
37 (56.9%)
0.5 (0.2–1.1)
29 (52.7%)
0.5 (0.2– 1.1)
0.046
Live birth
34 (61.8%)
Reference
39 (65.0%)
1.1 (0.5–2.6)
31 (47.7%)
0.6 (0.3–1.2)
25 (45.5%)
0.6 (0.3– 1.2)
0.06
2764 (2215– 3312)
0.26
Birthweight (g)**
3058 (2741–3375)
2994 (2638–3349)
3048 (2755–3340)
*Results are adjusted for donor age, recipient age, recipient BMI, male factor, number of prior cycles in donor, donor antral follicle count, donor status (known versus anonymous). Birthweight is adjusted for multiple pregnancy in addition to other covariates. **Lightest infant included if multiple pregnancy.
birthweight among BMI quartiles. There was a statistically significant decreased odds of a clinical pregnancy (P-trend ¼ 0.046) with increasing or BMI quartile (Fig. 1). There was also a trend toward lower odds of live birth with increasing BMI but this did not reach statistical significance (P-trend ¼ 0.06) (Fig. 1). Following sub-analysis which included only anonymous donors (Table IV and Fig. 1), there was a statistically significant decrease in the odds of clinical pregnancy (P-trend ¼ 0.02), and live birth (P-trend ¼ 0.03) with increasing BMI quartile. This significance did not change when the five patients who used a GnRH antagonist or flare protocol were omitted. A further sub-analysis of anonymous donors excluding recipients with BMI ≥25 showed the same trend toward lower
odds of clinical pregnancy and live birth with increasing BMI but did not reach statistical significance (Supplementary Table SII). Implantation rates and birthweights were similar among quartiles for all donors and for anonymous donors only. Cycle outcomes were also analyzed using donor BMI as a continuous variable instead of by quartile. A one-unit (1 kg/m2) increase in BMI is associated with a 0.9 times lower odds of achieving a clinical pregnancy (OR 0.9 (0.80– 1.00), P ¼ 0.049). In addition, a one-unit (1 kg/m2) increase in BMI is associated with a 0.9 times lower odds of achieving a live birth (OR 0.9 (0.82–1.02), P ¼ 0.09, however this did not reach statistical significance.
389
Oocyte donor BMI and reproductive outcomes
Figure 1 Adjusted odds ratios of clinical pregnancy and live birth by BMI quartile.
Table IV Cycle outcome by donor BMI, adjusted* OR (95% CI) or adjusted* rate (95% CI)—anonymous donors only. BMI Q1
BMI Q2
BMI Q3
BMI Q4
P-trend
............................................................................................................................................................................................. Cycles (number)
52
49
57
42
–
Biochemical pregnancy
Reference
0.8 (0.3–1.9)
0.6 (0.3–1.4)
0.4 (0.2– 1.0)
0.03
Clinical pregnancy
Reference
0.7 (0.3–1.7)
0.5 (0.2–1.1)
0.4 (0.1– 0.9)
0.02
Live birth
Reference
0.9 (0.4–2.2)
0.5 (0.3–1.2)
0.4 (0.2– 1.1)
0.03
*Results are adjusted for donor age, recipient age, recipient BMI, male factor, number of prior cycles in donor, donor antral follicle count, donor status (known versus unknown).
The results of a repeat analysis adjusting for total number of eggs retrieved (in addition to all other covariates previously controlled for including recipient BMI), indicated that a decrease in clinical pregnancy rate with increasing BMI persisted (P-trend ¼ 0.07 for all donors, P-trend ¼ 0.04 when excluding known donors). The trend toward lower live birth rate with increasing BMI also persisted when controlling for number of eggs retrieved (P-trend ¼ 0.06 for all donors, P-trend ¼ 0.047 when excluding known donors).
Discussion In this retrospective cohort study, increasing donor BMI was associated with lower clinical pregnancy and live birth rates. Given the nature of this donor-recipient model, which controlled for recipient BMI, the findings of this study suggest an upstream effect of donor BMI prior to conception. To our knowledge, this is the first study using a donor-oocyte model that
accounts for the BMI of the oocyte donor and recipient while studying the impact of BMI on clinical pregnancy and live birth outcomes. The observations of this study lend support to the hypothesis in animal models that there may be an effect of preconception BMI on the oocyte and subsequent IVF cycle outcomes. In studies using the oocyte donation model, the association of BMI and IVF cycle outcomes is inconsistent. Several studies utilizing this model report no impact of BMI on clinical pregnancy rates (Bellver et al., 2003; Styne-Gross et al., 2005), while the majority have reported a decrease in pregnancy rates (Bellver et al., 2007, 2013; Dessolle et al., 2009; DeUgarte et al., 2010; Luke et al., 2011) and live birth rates (DeUgarte et al., 2010; Luke et al., 2011; Bellver et al., 2013) associated with increasing BMI. These variable results are difficult to interpret as each of these studies varies greatly in terms of size and methodology. However, adjustments for the BMI of the oocyte donor were not employed in these analyses. A systematic review and meta-analysis of
390 outcomes in obese donor oocyte recipients found no association between obesity and chance of implantation, pregnancy or live birth and noted that ‘oocyte quality rather than endometrial receptivity may be the overriding factor influencing IVF outcomes in obese women using autologous oocytes’ (Jungheim et al., 2013). The authors further acknowledge that only recipient BMI was collected and accounted for, and that donor BMI may in fact affect oocyte quality, and therefore could have biased the results of the meta-analysis (Jungheim et al., 2013). Therefore, the oocyte donation model, which adjusts for both donor and recipient BMI, of this study may provide a more clinically relevant approach to delineate factors impacting outcomes in donor oocyte IVF cycles. An intriguing finding was the positive correlation between donor and recipient BMI, even when restricting to only anonymous cycles. One plausible explanation is that recipients emphasize having a donor with physical characteristics similar to their own over a donor with the healthiest possible BMI. In this study, peak serum estradiol levels declined with increasing BMI, and may be explained by the declining numbers of oocytes aspirated with increasing BMI. Since this finding is consistent with existing literature (Zhang et al., 2010), it could be assumed that the decrease in clinical pregnancy rate and live birth rate with increasing BMI is a reflection of the smaller oocyte cohort associated with increasing BMI in this study. However, following adjustment for the number of oocytes retrieved (while still adjusting for recipient BMI and all other covariates), the observation of a decline in clinical pregnancy rate and live birth rate with increasing BMI persisted. These findings imply that these IVF cycle outcomes are independent of oocyte quantity, and may be most impacted by oocyte quality. It is important to note that while cycle cancelations after stimulation start could be associated with BMI, the numbers were too small to assess for an association. These cycles are included as negative pregnancy outcomes to ensure that if BMI is associated with cycle cancelation, the pregnancy and live birth outcomes will not be biased. It is also noteworthy that cycle outcomes such as clinical pregnancy rate and live birth rate, which trend toward or reach significance with all donors included, became statistically significant when known donors were excluded. One explanation for this finding is that known donors are not as rigorously screened (as evidenced by their older age, higher BMI and higher day 3 FSH compared with anonymous donors), and may have other underlying health issues and comorbidities. Since the overall trend toward worse IVF outcomes with higher donor BMI became more obvious with exclusion of known donors, results within the anonymous donor population may further delineate the impact of BMI in the ‘ideal’ donor population. Furthermore, the persistent trend toward lower odds of clinical pregnancy and live birth with increasing BMI seen in anonymous donors when excluding recipients with BMI ≥25 further emphasizes the impact of donor BMI, as the lack of statistical significance may reflect the very small sample size of this subpopulation. When donor BMI was evaluated as a continuous variable, the finding that the odds of clinical pregnancy and live birth declined with increasing single units of BMI (1 kg/m2) is particularly intriguing, as this highlights the sensitivity of these outcomes to particularly small changes in BMI. The lack of statistical significance seen with live birth is likely a reflection of the relative clinical insignificance of a single-unit change in BMI. However, this also highlights the value of simultaneously examining
Cardozo et al.
data using donor BMI quartiles, as statistical significance is seen in declining odds of both clinical pregnancy and live birth with larger units of increase in donor BMI, increases which are also likely to be more clinically applicable. The findings of this study are consistent with existing animal literature that suggests an impact of preconception obesity on reproductive outcomes. The altered maternal metabolic environment of obesity likely exerts a long-term impact on the oocyte and subsequent postfertilization embryo, including delayed embryonic development, and the effect may persist when the oocyte or embryo is transferred to a metabolically normal host (Minge et al., 2008; Robker, 2008; Wyman et al., 2008; Luzzo et al., 2012). Several mechanisms have been proposed to explain the impact of obesity on the preconception oocyte, with altered mitochondrial activity thought to be the primary underlying mechanism leading to poor oocyte quality and inferior reproductive outcomes. Since mitochondria are maternally inherited, do not begin to replicate until after implantation, play a major role in early embryonic development, and obesity results in abnormal mitochondria in oocytes and embryos (Igosheva et al., 2010), it has been proposed that mitochondrial dysfunction may impede early embryonic development (Cummins, 2004; Dumollard et al., 2007; Van Blerkom, 2008, 2011; Cardozo et al., 2011). Thus, reduced oocyte quality could result in downstream alterations of embryo quality and implantation potential, which may not be reflected by early embryo morphology prior to transfer. An alternate hypothesis of obesity’s impact on oocyte quality prior to conception involves substantially increased spindle and chromosome alignment defects in oocytes of obese mice compared with controls, which are thought to result in embryos with ‘massive aneuploidy’ (Luzzo et al., 2012). The primary strength of this study is that our analysis models account for both donor and recipient BMI, and provide a comprehensive assessment of the impact of BMI prior to and subsequent to conception on reproductive outcomes. Moreover, the consistent use of protocols for donors and recipients was ideal to minimize variations in response. A limitation of this study is the relatively small sample size. Many of the results that are of borderline statistical significance demonstrate an notable trend but likely represent the relatively small sample size, and therefore larger future studies are needed. While the sample size is small, to our knowledge, this study still remains the largest to utilize the donor-recipient model to assess the impact of donor BMI, while controlling for recipient BMI, on IVF outcomes including clinical pregnancy and live birth rates. These findings provide a foundation for future research with larger numbers. Future larger studies may allow for the application of ROC curves based on donor BMI and clinical pregnancy and live birth rates, which could be used to determine an optimal cutpoint at which to consider limiting donors. An ideal future study would be an IRB approved randomized controlled trial using oocyte donors of variable BMI with normoweight recipients in order to compare IVF outcomes. However, given the significant emotional and financial burden of recipients undergoing oocyte donation cycles and desire to use the donor of their selection, enrollment and willingness to undergoing randomization may be a limiting factor. Since this study was completed at a single IVF program, some results may be reflective of treatment protocols and embryology laboratory techniques specific to our program. Future investigation should further explore the impact of small fluctuations in BMI on reproductive outcomes. A particularly intriguing future application of this study’s model would be to assess the association of
Oocyte donor BMI and reproductive outcomes
preconception weight loss on these same outcomes, and to investigate if the negative association of obesity on oocyte quality may be reversed with weight loss. How or if the findings of this study apply to women not undergoing IVF remains uncertain. Moreover, these data should not be interpreted to mean that all oocyte donors should be restricted to a BMI of less than 21.2 kg/m2. In summary, following simultaneous adjustments for oocyte donor and recipient BMI, we observed that increasing oocyte donor BMI is associated with worse IVF outcomes in donor-recipient IVF cycles, specifically a lower likelihood of clinical pregnancy and live birth. These findings imply that the negative impact of increasing BMI on IVF outcomes may occur at the level of the oocyte prior to fertilization. These observations are among the first in human studies to support the findings of prior animal studies, and warrant future investigation.
Supplementary data Supplementary data are available at http://humrep.oxfordjournals.org/.
Acknowledgements We would like to express our gratitude to the Deborah Kelly Center for Outcomes Research of the Vincent Department of Obstetrics and Gynecology, Massachusetts General Hospital for generous support of this study.
Authors’ roles E.R.C.: study conception and design, data acquisition, drafting of manuscript, critically revising manuscript, and final approval of manuscript. A.E.K.: study conception and design, data analysis, interpretation of data, critically revising manuscript, and final approval of manuscript. J.G.: data acquisition, critically revising manuscript, and final approval of manuscript. J.C.P.: study conception and design, interpretation of data, critically revising manuscript, and final approval of manuscript. A.K.S.: study conception and design, interpretation of data, critically revising manuscript, and final approval of manuscript.
Funding No external funding was used for this study.
Conflict of interest None declared.
References Barton SE, Missmer SA, Ashby RK, Ginsburg ES. Multivariate analysis of the association between oocyte donor characteristics, including basal follicle stimulating hormone (FSH) and age, and IVF cycle outcomes. Fertil Steril 2010;94:1292 – 1295. Bellver J, Rossal LP, Bosch E, Zuniga A, Corona JT, Melendez F, Gomez E, Simon C, Remohi J, Pellicer A. Obesity and the risk of spontaneous abortion after oocyte donation. Fertil Steril 2003;79:1136 – 1140.
391 Bellver J, Melo MA, Bosch E, Serra V, Remohi J, Pellicer A. Obesity and poor reproductive outcome: the potential role of the endometrium. Fertil Steril 2007;88:446 – 451. Bellver J, Pellicer A, Garcia-Velasco JA, Ballesteros A, Remohi J, Meseguer M. Obesity reduces uterine receptivity: clinical experience from 9,587 first cycles of ovum donation with normal weight donors. Fertil Steril 2013; 100:1050– 1058. Cano F, Garcia-Velasco JA, Millet A, Remohi J, Simon C, Pellicer A. Oocyte quality in polycystic ovaries revisited: identification of a particular subgroup of women. J Assist Reprod Genet 1997;14:254 – 261. Cardozo E, Pavone ME, Hirshfeld-Cytron JE. Metabolic syndrome and oocyte quality. Trends Endocrinol Metab 2011;22:103 – 109. Cardozo ER, Neff LM, Brocks ME, Ekpo GE, Dune TJ, Barnes RB, Marsh EE. Infertility patients’ knowledge of the effects of obesity on reproductive health outcomes. Am J Obstet Gynecol 2012;207:509 e501 –509 e510. Cardozo ER, Dune TJ, Neff LM, Brocks ME, Ekpo GE, Barnes RB, Marsh EE. Knowledge of obesity and its impact on reproductive health outcomes among urban women. J Community Health 2013;38:261 – 267. Cardozo ER, Thomson AP, Karmon AE, Dickinson KA, Wright DL, Sabatini ME. Ovarian stimulation and in-vitro fertilization outcomes of cancer patients undergoing fertility preservation compared to age matched controls: a 17-year experience. J Assist Reprod Genet 2015;32:587–596. Cummins JM. The role of mitochondria in the establishment of oocyte functional competence. Eur J Obstet Gynecol Reprod Biol 2004;115(Suppl 1):S23–S29. Dessolle L, Darai E, Cornet D, Rouzier R, Coutant C, Mandelbaum J, Antoine JM. Determinants of pregnancy rate in the donor oocyte model: a multivariate analysis of 450 frozen-thawed embryo transfers. Hum Reprod 2009;24:3082 – 3089. DeUgarte DA, DeUgarte CM, Sahakian V. Surrogate obesity negatively impacts pregnancy rates in third-party reproduction. Fertil Steril 2010; 93:1008– 1010. Dumollard R, Duchen M, Carroll J. The role of mitochondrial function in the oocyte and embryo. Curr Top Dev Biol 2007;77:21 – 49. Fitzmaurice GM, Laird N, Ware JH. Marginal Models: Generalized Estimating Equations (GEE): Applied Longitudinal Analysis. Hoboken, NJ: Wiley & Sons, 2004, 291– 321. Igosheva N, Abramov AY, Poston L, Eckert JJ, Fleming TP, Duchen MR, McConnell J. Maternal diet-induced obesity alters mitochondrial activity and redox status in mouse oocytes and zygotes. PLoS One 2010;5:e10074. Jungheim ES, Schon SB, Schulte MB, DeUgarte DA, Fowler SA, Tuuli MG. IVF outcomes in obese donor oocyte recipients: a systematic review and meta-analysis. Hum Reprod 2013;28:2720 – 2727. Loveland JB, McClamrock HD, Malinow AM, Sharara FI. Increased body mass index has a deleterious effect on in vitro fertilization outcome. J Assist Reprod Genet 2001;18:382– 386. Luke B, Brown MB, Stern JE, Missmer SA, Fujimoto VY, Leach R, Group SW. Female obesity adversely affects assisted reproductive technology (ART) pregnancy and live birth rates. Hum Reprod 2011;26:245– 252. Luzzo KM, Wang Q, Purcell SH, Chi M, Jimenez PT, Grindler N, Schedl T, Moley KH. High fat diet induced developmental defects in the mouse: oocyte meiotic aneuploidy and fetal growth retardation/brain defects. PLoS One 2012;7:e49217. Maheshwari A, Stofberg L, Bhattacharya S. Effect of overweight and obesity on assisted reproductive technology—a systematic review. Hum Reprod Update 2007;13:433 – 444. Minge CE, Bennett BD, Norman RJ, Robker RL. Peroxisome proliferatoractivated receptor-gamma agonist rosiglitazone reverses the adverse effects of diet-induced obesity on oocyte quality. Endocrinology 2008; 149:2646– 2656. Moragianni VA, Jones SM, Ryley DA. The effect of body mass index on the outcomes of first assisted reproductive technology cycles. Fertil Steril 2012;98:102 – 108.
392 Practice Committee of American Society for Reproductive Medicine. Obesity and reproduction: an educational bulletin. Fertil Steril 2008; 90:S21 – S29. Practice Committee of American Society for Reproductive Medicine; Practice Committee of Society for Assisted Reproductive Technology. Recommendations for gamete and embryo donation: a committee opinion. Fertil Steril 2013;99:47 – 62. Rich-Edwards JW, Goldman MB, Willett WC, Hunter DJ, Stampfer MJ, Colditz GA, Manson JE. Adolescent body mass index and infertility caused by ovulatory disorder. Am J Obstet Gynecol 1994;171: 171 – 177. Rittenberg V, Seshadri S, Sunkara SK, Sobaleva S, Oteng-Ntim E, El-Toukhy T. Effect of body mass index on IVF treatment outcome: an updated systematic review and meta-analysis. Reprod Biomed Online 2011;23: 421 – 439. Robker RL. Evidence that obesity alters the quality of oocytes and embryos. Pathophysiology 2008;15:115 – 121. Styer AK, Wright DL, Wolkovich AM, Veiga C, Toth TL. Single-blastocyst transfer decreases twin gestation without affecting pregnancy outcome. Fertil Steril 2008;89:1702– 1708. Styer AK, Gaskins AJ, Brady PC, Sluss PM, Chavarro JE, Hauser RB, Toth TL. Dynamic antimullerian hormone levels during controlled ovarian hyperstimulation predict in vitro fertilization response and pregnancy outcomes. Fertil Steril 2015;104:1153 –1161 e1157. Styne-Gross A, Elkind-Hirsch K, Scott RT Jr. Obesity does not impact implantation rates or pregnancy outcome in women attempting conception through oocyte donation. Fertil Steril 2005;83:1629–1634.
Cardozo et al.
Van Blerkom J. Mitochondria as regulatory forces in oocytes, preimplantation embryos and stem cells. Reprod Biomed Online 2008;16:553 – 569. Van Blerkom J. Mitochondrial function in the human oocyte and embryo and their role in developmental competence. Mitochondrion 2011;11:797–813. Vrontikis A, Chang PL, Kovacs P, Lindheim SR. Antral follicle counts (AFC) predict ovarian response and pregnancy outcomes in oocyte donation cycles. J Assist Reprod Genet 2010;27:383 –389. Wang JX, Davies M, Norman RJ. Body mass and probability of pregnancy during assisted reproduction treatment: retrospective study. BMJ 2000; 321:1320– 1321. Wattanakumtornkul S, Damario MA, Stevens Hall SA, Thornhill AR, Tummon IS. Body mass index and uterine receptivity in the oocyte donation model. Fertil Steril 2003;80:336 – 340. Wyman A, Pinto AB, Sheridan R, Moley KH. One-cell zygote transfer from diabetic to nondiabetic mouse results in congenital malformations and growth retardation in offspring. Endocrinology 2008;149:466 – 469. Yeh JS, Steward RG, Dude AM, Shah AA, Goldfarb JM, Muasher SJ. Pregnancy outcomes decline in recipients over age 44: an analysis of 27,959 fresh donor oocyte in vitro fertilization cycles from the Society for Assisted Reproductive Technology. Fertil Steril 2014;101:1331– 1336. Zenke U, Chetkowski RJ. Transfer and uterine factors are the major recipient-related determinants of success with donor eggs. Fertil Steril 2004;82:850 – 856. Zhang D, Zhu Y, Gao H, Zhou B, Zhang R, Wang T, Ding G, Qu F, Huang H, Lu X. Overweight and obesity negatively affect the outcomes of ovarian stimulation and in vitro fertilisation: a cohort study of 2628 Chinese women. Gynecol Endocrinol 2010;26:325 – 332.
