Document Type : Original Article
Authors
1 Department of Biostatistics, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran
2 Department of Epidemiology and Reproductive Health, Reproductive Epidemiology Research Center, Royan Institute for Reproductive Biomedicine, ACECR, Tehran, Iran
Abstract
Keywords
In Iran, the average rate of infertility, primary and secondary
infertility and current infertility is estimated to be
10.9% [95% confidence intervals (CI): 7.4-14.4], 10.6%
(95% CI: 5.3-16.0), 2.7% (95% CI: 1.9-3.5) and 3.3%
(95% CI: 2.7-3.8), respectively (
An
Most studies on ART data have only inspected a part of
infertile women’s data (
In the GEE analysis it is assumed that, the response is independent
from the number of observations in the cluster
(the cluster size) (
This historical cohort study includes 996 cycles of 511 infertile women who were enrolled in ART treatments between April 2011 and March 2012 in Royan institute, Iran. Only women who experienced embryo transfer were eligible to be included in the present analysis. All variables in this study were defined based data extracted from the medical record of the individuals, by trained nurses. The outcome variable was success or failure at four stages: i. Chemical pregnancy [a transient increase in serum beta-human chorionic gonadotropin (ß-hCG)], ii. Clinical pregnancy (presence of an intrauterine gestational sac), iii. Spontaneous abortion (pregnancy loss before 20 completed weeks of pregnancy), and iv. Delivery (live birth of at least one baby).
Cycles resulted in failure types other than the four above-mentioned ones, were excluded from the study and couples who required donation or gestational carrier, were not eligible for enrollment. Covariates considered in this study were women’s age (under 35, 35 to 37, 38 to 40, above 40), type of cycle (fresh or frozen embryo transfer), the number of embryos transferred and having polycystic ovarian syndrome (PCOS) during IVF cycles. Some other measured covariates were woman-specific, such as age at the first cycle while some others were cycle-specific, such as type of cycle or the number of embryos transferred.
The study was approved by the Ethics board of research of Royan institute (Ethical code: EC/90/1086). Informed consent was obtained from all subjects when they intended to start the treatment. Subjects were assured that the results would be published following statistical evaluations and no personal data would be disclosed.
The outcome at each stage (chemical pregnancy, clinical
pregnancy, spontaneous abortion (SAB) and delivery)
was considered as the binary response variable representing
the success or failure of the stage. The probability of
success occurrence at a specific stage of ART cycle, could
be associated with the stage, cycle number, and covariates
of interest. The main challenge is considering the correlations
among repeated cycles of each woman, as well as
correlations among the outcomes of multiple stages within
each cycle. To consider these correlations, GEEs were used
according to the model presented by Maity et al. (
This study includes 511 women with a total of 996 IVF
cycles, each woman having 1-3 cycles leading to embryo
transfer. The mean (SD) age of women was 35.75 (5.12)
years old and 86 (16.8%) of women had PCOS. Among the
cycles included in this study, 585 (59%) were cycles with
fresh embryo transfer and the median (inter quartile range)
of the number of embryos ready for transfer was 3 (
Since the number of cycles that each woman experienced
is reversely associated with the success/failure
at different stages, conditional on other predictors (OR:
0.68, 95% CI: 0.52-0.89, P=0.005), cluster size is believed
to be informative and CWGEE has been suggested
for handling this situation (
GEE and CWGEE models used in this study incorporated
the data from repeated IVF cycles and multiple stages,
with a separate intercepts for stage (
To explore the differing effect of fresh and frozen embryo
transfer on the odds of success at various stages, the
interaction term between type of embryo(s) transferred
and failure type was included in the model. Although
women receiving fresh embryo transfer showed significantly
better results in clinical pregnancy, from then on,
women receiving frozen embryo transfer could successfully
continue in the same way as those receiving fresh
embryos (
Relationship between IVF outcomes and IVF/participants characteristics
IVF and participants characteristics | Unweighted GEE | Cluster weighted GEE | ||
---|---|---|---|---|
OR (95% CI) | P value | OR (95% CI) | P value | |
Intercepts | ||||
Chemical pregnancy | 1 ( reference) | - | 1 ( reference) | - |
Clinical pregnancy | 2.11 (2.08, 2.15) | <0.001 | 2.12 (2.09, 2.18) | <0.001 |
SAB | 2.20 (2.13, 2.29) | <0.001 | 2.22 (2.14, 2.34) | <0.001 |
Delivery | 6.43 (3.42, 15.76) | 0.010 | 8.69 (3.59, 30.03) | 0.009 |
Embryos transferred number | 1.18 (1.01, 1.38) | 0.031 | 1.21 (1.03, 1.42) | 0.021 |
PCOS | ||||
Yes | 0.74 (0.52, 1.06) | 0.102 | 0.75 (0.52, 1.10) | 0.138 |
No | 1 ( reference) | - | 1 ( reference) | - |
Type of embryo(s) transferred | ||||
Fresh | 1 ( reference) | - | 1 ( reference) | - |
Frozen | 2.50 (1.87, 3.35) | <0.001 | 2.26 (1.66, 3.07) | <0.001 |
Age categories (Y) | ||||
<35 | 1 ( reference) | - | 1 ( reference) | - |
35-37 | 0.86 (0.38, 1.28) | 0.460 | 0.87 (0.57, 1.31) | 0.504 |
38-40 | 0.68 (0.46, 1.00) | 0.052 | 0.67 (0.45, 1.00) | 0.050 |
>40 | 0.74 (0.51, 1.07) | 0.109 | 0.76 (0.52, 1.11) | 0.161 |
IVF;
Log odds of success at multiple points during the IVF cycle with 95% confidence intervals.
IVF;
There are some existing approaches to model IVF data including multiple cycles with multiple failure types (
Since the number of cycles experienced by each infertile woman is believed to be associated with the success/failure of IVF outcome, studies on these type of data involve informative cluster size and GEE and CWGEE, might show different results as GEE assumes that cluster size is non-informative. This historical cohort study on Iranian infertile women also demonstrated strong reverse associations between the number of cycles and odds of success in IVF outcomes, indicating the presence of informative cluster size (
Based on both GEE and CWGEE, our results also suggest that successful IVF outcomes seem to be associated with performing frozen embryo transfer compared to fresh embryo transfer. This could be explained by the fact that the endometrium is more receptive in frozen embryo transfer during the endometrial priming than in fresh embryo cycles; therefore, frozen embryo cycles could lead to a better embryo-endometrium synchrony (
In our study, having PCOS was not significantly associated
with odds of success in IVF procedures in either of
the models which was not consistent with some previous
research that found that women with PCOS have an increased
prevalence of miscarriage, both after spontaneous
and induced ovulation (
A great deal of previous research has indicated significant
associations between women age and fertility (
In this study, data from repeated IVF cycles was used by including the correlation among them; however, not including some variables of couples undergoing IVF, such as pretreatment variables, embryo quality, oocyte and sperm quality and also stimulation and laboratory variables is a limitation of this study. Data on previous cycles, which infertile women might have undergone in other infertility centers, was not included in this study due to lack of a national registry.
Frozen embryo transfer was positively associated with odds of success compared to cycles with fresh embryo transfer; but, cycles with fresh embryo transfer had better results in clinical pregnancy compared to frozen embryo transfer. The number of embryos transferred and women’s age were significantly associated with odds of success.