Hum. Reprod. Advance Access originally published online on November 9, 2006
Human Reproduction 2007 22(2):548-557; doi:10.1093/humrep/del403
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Construction of an evidence-based integrated morphology cleavage embryo score for implantation potential of embryos scored and transferred on day 2 after oocyte retrieval
1 Carl von Linné Clinic 2 Uppsala Clinical Research Center, Uppsala University, Uppsala Science Park, Uppsala, Sweden and 3 Department of Gynaecological and Obstetrical Sciences, University of Turin, Turin, Italy
4 To whom correspondence should be addressed at: Carl von Linné Clinic, Uppsala Science Park, S-751 83 Uppsala, Sweden. E-mail: jan.holte{at}linne.se
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Evidence-based morphological embryo scoring models for ranking of implantation potential are still scarce, and the need for a precise model increases when aiming for singleton pregnancies. METHODS: Prospectively, 2266 IVF/ICSI double-embryo, day 2 transfers were studied. The five variables scored in 3- to 5-step scales for the embryos transferred are blastomere number (BL), fragmentation, blastomere size variation (‘equality’, EQ), symmetry of the cleavage and mononuclearity in the blastomeres (NU). The scoring results of embryos with an individual traceability from scoring to implantation, i.e. treatments resulting in either no implantation (n = 1385) or twin implantation (n = 228), were studied for prognostic potential. RESULTS: Although all five variables correlated highly with implantation potential, only BL, NU and EQ remained independently significant after regression analysis. The equation thus derived formed the basis for a 10-point integrated morphology cleavage (IMC) embryo score. A table with the scoring point for each possible combination of the embryo variables is presented. The scoring model was statistically validated on the singleton pregnancy group (n = 653). CONCLUSIONS: We suggest that this IMC embryo scoring, incorporating cleavage stage and information on the variation in blastomere size and the number of mononucleated blastomeres, may optimize embryo ranking and selection for day 2 transfers.
Key words: embryo morphology/embryo score/implantation potential/prediction
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Selection of embryos to transfer with the greatest implantation possibility is largely based on clinical tradition and less derived from evidence-based medicine. Although there is general agreement among embryologists as to what morphological features are characteristic of a top embryo in the cleavage stage, evidence is still lacking for the ranking of implantation potential of non-top embryos. The need for establishing a greater knowledge about embryo quality variables and thus constructing reliable scoring systems is becoming increasingly more evident. Several studies show that embryo quality is one of the most important prognostic factors for pregnancy chance in IVF apart from age (Puissant et al., 1987
; Staessen et al., 1992; Steer et al., 1992; Shulman et al., 1993; Giorgetti et al., 1995; Van Royen et al., 1999; Hardarson et al., 2001; Terriou et al., 2001; Hunault et al., 2002). The increased obstetric and perinatal risks involved in multiple pregnancies (Bergh et al., 1999; Stromberg et al., 2002) urge the clinician to aim at singleton pregnancies, which increases the demands on knowledge of the implantation potential of the individual embryo. Other reasons for establishing evidence-based embryo quality criteria are the needs for providing individualized prognostic information to the couple, for permitting comparison of embryo quality between treatment cycles, for the evaluation of different culture conditions and finally to allow standardizing embryo grading between embryologists.
The reason for the lacking scientific data is largely the difficulties in following the fate of an individual embryo. The dominating clinical practice of transferring more than one embryo makes deduction from embryo quality variables unreliable, when the resulting pregnancy contains fewer sacs than the number of transferred embryos. Therefore, the available scientific data to date are based on studies containing limited number of treatments with a traceable association between embryo and implantation (Giorgetti et al., 1995; Ziebe et al., 1997; Van Royen et al., 1999, 2001). Nevertheless, such studies have provided important information on morphological factors of importance for implantation potential. Among such factors, cleavage rate clearly provides powerful prognostic information (Giorgetti et al., 1995; Ziebe et al., 1997; Van Royen et al., 1999, 2001). Severe fragmentation of the embryo is associated with poor prognosis (Giorgetti et al., 1995; Ziebe et al., 1997; Van Royen et al., 1999, 2001). Other variables, such as visible mononuclearity (Palmstierna et al., 1998; Saldeen and Sundstrom, 2005) and absence of multinuclearity within the blastomere (Ziebe et al., 1997; Van Royen et al., 1999, 2001, 2003), uniformity in blastomere size (Giorgetti et al., 1995; Ziebe et al., 1997, 2003; Hardarson et al., 2001), symmetry of the cleavage (Dor et al., 1986) and variation in zona thickness (Garside et al., 1997; Palmstierna et al., 1998; Gabrielsen et al., 2001), also seem to have prognostic power. However, many variables are interrelated, making it uncertain which factors have independent prognostic impact. To construct a reliable embryo score, which incorporates only independently significant embryo variables and in correct proportions, we must test all factors of univariate association with implantation potential against each other in a statistical multiple regression procedure. Large prospective studies utilizing this statistical method to construct embryo scoring systems are still lacking in the literature.
This study was designed to investigate the impact of cleavage stage and four morphological embryo variables for implantation potential of the individual embryo in day 2 transfer procedures. Only embryo features possible to record with relative ease during routine real-time investigations were included. Therefore, among the variables described, zona thickness variation was omitted because of the necessity for video recording and, in reality, retrospective analyses for using this variable (Garside et al., 1997; Palmstierna et al., 1998; Gabrielsen et al., 2001). To allow the evaluation of each morphological feature as a continuous variable, which increases the statistical power, we constructed novel graded scoring systems for the variables blastomere size variation, symmetry of the cleavage and the occurrence of blastomeres with a visible single nucleus. The degree of fragmentation was measured according to established scoring procedures. Apart from these morphological features of the embryo, the cleavage rate was recorded (number of blastomeres) as a fifth parameter. The primary end-points were each variable’s correlation to implantation and the predictive power of all variables in aggregate, thus resulting in an embryo scoring model. Two-embryo transfers leading to either twin or failed implantation, i.e. individual embryos with a known implantation potential, formed the primary study group. The derived prognostic model, an integrated morphology cleavage (IMC) scoring, was subsequently statistically validated on the remaining treatments in the study population, i.e. two-embryo transfers leading to a single implantation.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
All (n = 2266) IVF (n = 1338) or ICSI (n = 928) treatments at the Carl von Linné Clinic during a 3-year period that resulted in the transfer of two embryos on day 2 after oocyte retrieval were studied. During this period, 1999–2001, double-embryo transfer was totally dominant in Sweden, and thus, the material amounts to 91% of the treatments at the clinic during this time interval. All ovarian stimulations were conducted with recombinant FSH in 98.5% of the cycles during long GnRH-agonist down-regulation (in 35 protocols, GnRH antagonist was used). The mean age of the women was 33.8 ± 4.3 years (SD). Oocytes were aspirated 35–37 h after hCG injection, inseminated (or injected with sperm after denuadation for ICSI) after 2–6 h of incubation and cultured in IVF medium (Medicult, Denmark) in a 5% CO2 incubator at 37°C. Fertilization was checked 16–20 h after the insemination. The embryos were evaluated on day 2 after oocyte retrieval; scoring and selection for transfer were performed immediately before embryo transfer. The Petri dish with the embryos was placed on the heating plate under a stereomicroscope (Nikon SMZ 1500) to maximal magnification (x110), and the embryos were gently rolled with a Pasteur pipette to assess the five morphological parameters. Light was directed obliquely through an angled mirror beneath the microscope to enable the visualization of the nuclei. This combination of intermediate magnification and the possibility to turn the embryos around under oblique light is in our experience a superior technique for the visualization of nuclei and to score for size variation of the blastomeres as well as symmetry of the cleavage. Results from a subsequent multicentre study support this notion (Holte, 2006
). Embryo transfer was performed at 48–52 h after oocyte retrieval. The following parameters were recorded for the embryos transferred: number of blastomeres (‘Blastomeres’, 2, 3, 4, 5 or 
6 blastomeres), degree of fragmentation (‘Fragmentation’), variation in blastomere sizes (‘Equality’), symmetry of the cleavage (‘Symmetry’) and the presence of a single nucleus within the blastomere (corrected for the number of blastomeres, ‘Nucleus’). The degree of fragmentation was scored as 0 (no fragments), 1 (
10% fragmentation), 2 (>10 25%), 3 (>25 50%) and 4 (>50% fragmentation). Equality was scored according to a three-level system, where 0 denotes uniform size of the blastomeres, 1 denotes varying size but <50% variation and 2 denotes >50% variation in blastomere size. Similarly, Symmetry was scored as 0, 1 and 2, describing full symmetry of the cleaved embryo, slightly asymmetric cleavage to pronounced asymmetry, respectively. The difference between these two parameters is thus that Equality scores variation in blastomere size, whereas Symmetry scores the arrangement of the blastomeres three dimensionally. The parameter Nucleus was defined as the number of visible mononucleated blastomeres divided by the total number of blastomeres in the embryo (to correct for cleavage rate); a ratio of 0–0.25 gives a Nucleus score of 0, >0.25–0.50 gives a Nucleus score of 1, >0.50–0.75 gives a score of 2 and a ratio of >0.75–1 equals a Nucleus score of 3. The score –1 denotes that the embryo contains at least one multinucleated blastomere. The scoring procedure is illustrated in Figure 1.
|
Because the exact fate, i.e. implantation potential, of each single transferred embryo could only be traced in twin pregnancies (ultrasonically identified the presence of two gestational sacs; Group 2, n = 228) or treatments with a negative pregnancy test (Group 0, n = 1385), these two subgroups formed the main study material (n = 1613). In the statistical analysis, each embryo in this main study group was treated as an individual record. Thus, 456 embryos in this material had a proven implantation, and 2770 had a proven failed implantation. As the singleton pregnancy group is not included in this data evaluation, all implantation figures are lower than the corresponding true implantation percentage for each variable. A correction factor of 1.731 (i.e. the ratio between 456/3226 and the general implantation of 1109/4532) is therefore used to extrapolate the implantation figures to be valid for the data in our entire population.
In a second step, the validity of the prediction model derived from the Group 0 and 2 material was tested on the singleton pregnancy group (‘Group 1’, n = 653).
Statistics
Use of Groups 0 and 2
Univariate relations between predictor variables and the outcome (Group 0 or 2) were described as cross-tabulations and statistically analysed in logistic regression models. Each predictor variable’s functional relation to the outcome was examined, and, if necessary, the variable was transformed. The interrelations between these predictor variables were examined using Spearman’s rank correlation coefficients. Thereafter, a conditional multiple logistic regression model, with forward selection of variables, was applied to find the predictor variables with independent impact on the outcome. Conditional regression means that the model takes into account the dependence between the two embryo records within the same treatment. In this model, P values from two-tailed tests <5% were considered statistically significant. Also, the clinical importance of each predictor variable’s effect was judged before the variable finally was considered worthwhile in the prediction model. From this model, estimated probabilities were calculated for each treatment using a weighed sum of mean embryo variable values where the estimated regression coefficients were used as weights. This calculated value is the estimated probability of belonging to Group 2, conditional upon being in Group 0 or 2. For each group, the mean estimated probability was presented. The relative merits of the predictor variables were illustrated by the calculation of the estimated probability for a number of combinations of the predictor values. The aim was to show the impact of each variable when the others were held constant (to their mean values) and to show the best and the worst combinations of predictor values and their associated probabilities.
Use of Group 1
The prediction model estimated from Groups 0 and 2 was applied to Group 1 by calculating estimated embryo scores using the predictor variable values from Group 1 and the estimated regression coefficients from Groups 0 and 2. The assumption was that a valid model would imply mean estimated embryo scores for Group 1 that were in between the mean estimated embryo scores from Groups 0 and 2. These three mean estimated embryo scores were presented with 95% confidence intervals (95% CIs). Group 1 data were also used for the construction of a model showing the impact of age and IMC scores on implantation possibility (Figure 3). In this calculation, Group 1 results were used according to the assumption that, in general, the embryo with the highest IMC score was most likely to implant. This assumption is strongly supported by the findings of the above-preformed analysis. These data were used together with Group 0 and 2 data. The reason for including Group 1 data was the lack of data in Group 2 in the extremes of the range for IMC scores (low) and the age groups (high).
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The total clinical pregnancy rate was 39% (881/2266), with a twin rate of 26% (228/881), corresponding to an implantation rate of 24.5% (1109/4532). There were no differences between IVF and ICSI results (implantation rates 25 versus 23%, not significant). Also, the results for morphological embryo variables did not differ significantly between IVF and ICSI. Therefore, all results are shown for the two methods in aggregate.
Morphological variables: analysis restricted to Groups 0 and 2
Figure 2 shows univariate implantation percentage for the five morphological variables, i.e. the relation between each variable and the implantation percentage. It also shows the distribution of the different scores for the five variables. All variables could best be described by linear functions except Blastomere, which was closer to a two-degree function, a relationship suggested by the fact that implantation figures increased from 2- to 4-cell embryos but declined, although less pronounced, when the embryo exhibited more than four blastomeres at transfer (Figure 2). Statistical correction for the exact time passed between insemination and transfer did not affect the results for any variable significantly (data not shown). Each variable showed a highly significant degree of correlation with the percentage of implantation, with a similar span in implantation figures between the lowest and the highest score in each variable (Figure 2).
|
The five variables showed varying degrees of significant intercorrelations (Table I). Blastomeres correlated only with fragmentation (inversely), whereas all the other variables showed significant intercorrelations.
|
After a multiple logistic regression analysis, the three variables Blastomeres, Equality and Nucleus remained significant predictors for implantation (Table II). The other two morphological variables, Fragmentation and Symmetry, resulted insignificant in the logistic regression analysis. The multiple model indicated the relative impact by Nucleus and Equality on treatment outcome to be of similar magnitude, whereas the impact of Blastomere varied depending on cleavage stage. Thus, the impact of a one-unit increase of Nucleus resulted in 1.42 times increase in implantation rate, and a one-unit decrease in Equality raised the implantation rate by 1.80 times. For Blastomere, the most pronounced impact on treatment outcome was observed for the step from three to four blastomeres, which was associated with an increase in implantation rate by 4.5 times.
|
The logistic regression analysis resulted in the following equation for the estimated probability of implantation: P = 1.731 x exp(g)/[1 + exp(g)], where g = –4.54 + 0.60 x Blastomeres – 1.04 x abs(Blastomeres-4) – 0.59 x Equality + 0.35 x Nucleus (Table II), where the non-linear function of Blastomere is described by Blastomere (number of blastomeres) and the term abs(Blastomeres-4), i.e. the deviation from the ideal number of blastomeres. (The combination of these two terms describing the variable Blastomere in the multiple model could thus distinguish the importance of, for example, Blastomere 5 from that of Blastomere 3 on implantation chance.) To construct the IMC score, we adapted the resulting figure of this equation to a 10-point scale (by multiplying with 1.53 and adding 11.64), in which a ‘top’ embryo with four blastomeres (Blastomere 4), perfect equality in size of the blastomeres (Equality 0) and a visible single nucleus in each blastomere (four observed nuclei divided by four blastomeres = Nucleus 3) gives a score of 10 points, whereas a 2-cell embryo (Blastomere 2) with large variation in blastomere size (Equality 2) and at least one multinucleated blastomere (Nucleus –1) would give a score of 1. Table III summarizes the embryo scores given for any combination of blastomere number, equality score and number of blastomeres with a visible nucleus.
|
Fit between IMC scores and observed implantation rates (Groups 0 and 2)
Table IV summarizes the number of observations in Groups 0 and 2 for IMC scores 1–10 (in full-digit groups) and the corresponding observed implantation rates, ranging from 0 to 40%, and resulting in a correlation between IMC scores and implantation rate of 0.93 (weighted Pearson’s correlation coefficient; P
0.0001). The figures in the lower IMC range should be interpreted with care because of the low number of observations. If plotted, the association between scores and implantation would describe a graph close to the graph shown for age group 32–35 years in Figure 3, as expected, given the mean age of the entire population studied (33.8 ± 4.3 SD).
|
Validation of the prediction model applied on Group 1
The derived prediction model was applied to the treatments resulting in a single implantation (Group 1). The validation was tested with three different alternatives: the highest embryo score, the lowest score and a mean of the two embryo scores. The results gave predicted mean embryo scores in between those for failed implantation (Group 0) and full implantation (Group 2), with increasing levels from the lowest to the highest embryo scores (Table V), indicating the model to be valid.
|
Impact of women’s age on the IMC embryo scoring model
Age did not show a significant statistical interaction with the IMC embryo score, i.e. the effect of age was similar over the entire age range. Implantation rates for different embryo scores in six age groups were calculated utilizing the entire study population to increase statistical power. Group 1 results were included in this calculation, based on the presumption that the embryo with the highest embryo score was more likely to implant. The implantation rates best fitted the following equation: 100 x exp(–4.5265 + 0.4621 x EmbSc – 0.2934 x AgeGr)/[1 + exp(–4.5265 + 0.4621 x EmbSc – 0.2934 x AgeGr)], where EmbSc is embryo score from 1 to 10 and AgeGr is age group from 1 to 6.
The resulting graphs are depicted in Figure 3. The impact of embryo score on implantation is proportionally affected by age over the entire range of embryo scores. For example, at IMC score 10, the implantation rate is 45 versus 15% in the youngest age group compared with the oldest group, i.e. 3 times higher. At score 8, the implantation rates are 24 and 7, respectively, again a ratio of similar magnitude.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The lack of an exact traceability from the individual embryo to implantation in IVF programmes with more than one embryo per transfer has long formed a major obstacle for designing powerful studies on the prognostic impact of various morphological embryo variables. Most previous embryo scoring models were therefore derived from retrospective analyses of limited number of treatments or mainly clinical observations and thus had the characteristics of aggregates of reasonable implantation predictors. Furthermore, previously published embryo scoring algorithms were not based on multivariate analyses, which is a prerequisite for a balanced integrated model, in which predicted implantation chance is exactly proportional to the score (Steer et al., 1992
; Giorgetti et al., 1995; Ziebe et al., 1997; Van Royen et al., 1999, 2001). Nevertheless, some models, such as the cumulative embryo score (CES) (Steer et al., 1992), have become widespread and gained appreciation as clinically uncomplicated prediction models with good accuracy for reducing high-order multiple pregnancies (Steer et al., 1992; Visser and Fourie, 1993; Copperman et al., 1995; Moon et al., 2000; Terriou et al., 2001; Child et al., 2002; Laasch and Puscheck, 2004). However, given the amounting evidence for increased fetal and obstetric risks also for twin pregnancies (Bergh et al., 1999; Stromberg et al., 2002), and not only for triplets and quadruplets, the difficulties in finding the right embryos and right patients for single-embryo transfer increase the demands for accurate embryo scoring models. The design of this study was unique in the sense that it was large in number and power, was prospective and allowed tracing of the individual embryo. Moreover, not only the number of blastomeres and the degree of fragmentation but also the three variables equality, mononuclearity in the blastomeres and symmetry of the cleavage were scored as semi-continuous in 3- to 5-step scales. This strategy allows graded information and increases the statistical power compared with the more commonly applied dichotomous scoring (‘even–non-even’, ‘visible nucleus–not visible nucleus’). Finally, the statistical procedure with multiple regression analysis is necessary to find out which variables have independent power and thus are not merely passively associated with one or several other variables and to find the correct power balance between such independent variables.
It could be argued that the ideal approach to study the morphological determinants of a single embryo’s implantation potential would be to analyse exclusively single-embryo transfers. However, in most single-embryo transfer programmes, only ‘top’ embryos are transferred, and thus, an optimal span in variables for statistical evaluation cannot be reached by this approach. Alternatively, data from treatments, which result in only a single embryo available to transfer, could be analysed. This has been done, indeed producing important information, but the evaluation of such data is hampered by the fact that these treatments mainly involve women with poor response, poor embryo quality and low implantation figures, thus again not allowing a wide span of morphological variation (Giorgetti et al., 1995). The strategy of our study was close to those applied in the retrospective analyses published by Ziebe et al. (1997) (cycles in which the two or three transferred embryos were of identical morphological score) and by Van Royen et al. (1999, 2001) (all pregnancies of equal implantation number as the number of embryos transferred). In the present study, the prospective design and the large number of embryos with a completely known fate (3226) made it possible to produce statistics with great power.
The five variables in the present study have never been tested together before, although for each variable, there are indications on its usefulness as a marker for embryo competence (Dor et al., 1986; Giorgetti et al., 1995; Ziebe et al., 1997, 2003; Palmstierna et al., 1998; Van Royen et al., 1999, 2001; Fisch et al., 2001; Hardarson et al., 2001; Saldeen and Sundstrom, 2005).
The cleavage rate turned out to be the most powerful marker for implantation potential, corroborating other reports on this variable’s importance (Giorgetti et al., 1995; Ziebe et al., 1997; Van Royen et al., 1999, 2001). The ideal cleavage rate on day 2 was four blastomeres, with a markedly lower implantation potential for embryos with a cleavage stage below that and, to a lesser degree, above that. This finding of an optimal cleavage rate is in general agreement with previous observations (Cummins et al., 1986; Claman et al., 1987; Staessen et al., 1992; Giorgetti et al., 1995; Ziebe et al., 1997; Van Royen et al., 1999; Magli et al., 2001). In contrast to the other variables tested in the study, the cleavage rate variable was thus not linear, which must be taken into account when using it in an embryo scoring model. This knowledge could therefore be used for modifying existing models in which the number of blastomeres is incorporated as a linear factor (Steer et al., 1992; Visser and Fourie, 1993; Copperman et al., 1995; Dean et al., 2000; Fisch et al., 2001; Child et al., 2002; Laasch and Puscheck, 2004) or as a dichotomous factor (Terriou et al., 2001).
The prognostic importance of equally sized blastomeres has previously been observed clinically (Giorgetti et al., 1995; Ziebe et al., 1997; Hardarson et al., 2001) and in studies showing an increased prevalence of chromosomal abnormalities in embryos with unevenly sized blastomeres (Hardarson et al., 2001; Magli et al., 2001; Ziebe et al., 2003). Previous investigators applied a dichotomous scoring system for blastomere size, and the independent importance of this variable has not been tested in multivariate analyses. In fact, many clinically used embryo scoring systems do not permit scoring of both blastomere size variation and fragmentation in a single embryo, because the system through its rigidity forces the embryologist to score for either one or the other variable, but not both simultaneously. In the present study, we made an attempt to grade the equality variable into three degrees and test its validity in competition with the other four variables. The highly significant linear correlation of the equality variable with implantation rate suggests that such scoring in three steps for blastomere size variation is useful, and this is further underlined by the variable’s qualification as the second most important independent variable for implantation potential.
According to the present results, visible mononuclearity in the blastomeres is another independent sign of embryo competence, qualifying as the third component in the IMC score. The strong negative impact on the implantation potential of multinucleated blastomeres, previously observed (Van Royen et al., 1999, 2001, 2003), was illustrated by our findings of multinuclearity as the least favourable marker in this variable, with a lower implantation potential than that observed for embryos with a complete absence of visible nuclei. The positive impact of visible mononucleated blastomeres was previously observed (Palmstierna et al., 1998) and has been recently corroborated (Saldeen and Sundstrom, 2005). Our findings suggest that implantation potential increases linearly with the number of blastomeres with a visible single nucleus and thus that this variable preferably should be recorded as a type of continuous variable, and not dichotomously, or indeed merely as the absence of multinuclearity. A previous study on a limited number of embryos also concluded that the number of mononucleated blastomeres scored higher than most conventional markers of embryo quality (Palmstierna et al., 1998). Indeed, the findings of that study suggested the nucleus variable to be superior to cleavage stage, although in that study, the nucleus variable was not corrected for the number of blastomeres. This makes it difficult to draw solid conclusions from the analysis, because such a variable then incorporates cleavage stage. More recently, another Swedish study found the presence of a single nucleus within each blastomere in 4-cell embryos to predict a higher implantation rate than if one or more blastomeres did not show a nucleus. Interestingly, similar to our results, degree of fragmentation was insignificant, whereas in contrast to our findings, equality of the blastomeres did also result insignificant in the multiple analysis. In that study, however, equality was scored only in two and not in three degrees. Furthermore, the material consisted exclusively of 4-cell embryos in elective single-embryo transfers and thus embryos of a generally higher quality and less morphological variation than in the present study (Saldeen and Sundstrom, 2005).
The degree of fragmentation and symmetry of the cleavage, although highly significant in univariate analyses, did not retain significance in the multiple model. This suggests that fragmentation degree, an established marker, could be substituted by scoring for blastomere size variation and visible single-cell nucleus when attempting for an optimal multivariate prognostic power. This might raise questions about the true biological relevance of slight-to-moderate fragmentation at the early cleavage stage for embryo competence. Several authors have previously observed that slight fragmentation has no negative impact on implantation (Giorgetti et al., 1995; Ziebe et al., 1997, 2003; Van Royen et al., 1999), and small fragments may disappear through lysis or resorption during culture (Hardarson et al., 2001; Van Blerkom et al., 2001). Fragmentation was not associated with chromosomal abnormality rate in a recent study (Ziebe et al., 2003). A superiority of cleavage stage over degree of fragmentation for judging embryo competence has also been suggested by several investigators (Giorgetti et al., 1995; Ziebe et al., 1997, 2003; Van Royen et al., 1999). Furthermore, in two studies when fragmentation was tested in a multivariate way, it lost significance in comparison with zona pellucida variation (Palmstierna et al., 1998) and the number of mononucleated blastomeres (Palmstierna et al., 1998; Saldeen and Sundstrom, 2005).
The lack of independent significance (although strongly significant univariately) for fragmentation in the present study should not, however, be interpreted as the evidence for a lack of biological significance, but rather that as a marker of embryo competence in the present scoring scale, it loses power when being challenged by a ‘new’ set of strong markers. It is possible that scoring of fragmentation should take into account also the localization of fragments (Alikani et al., 1999; Ebner et al., 2001; Van Blerkom et al., 2001) and that the occurrence of fragmentation at later cleavage stages has more biological relevance (Alikani et al., 1999; Ebner et al., 2003). Severe fragmentation was associated with increased occurrence of malformations (Ebner et al., 2001), which might suggest that an embryo model for birth of a healthy child, instead of implantation, which was the end-point in the present study, would upgrade the importance of fragmentation.
Symmetry of cleavage is the variable in this study with the least previous scientific support (Dor et al., 1986; Hill et al., 1989; Fisch et al., 2001), and earlier studies discussing embryo symmetry usually incorporated this variable with size variation of the blastomeres. It is probable that slight asymmetry is normal and has no negative impact when observed in an embryo at the process of cleaving, whereas it may otherwise signal increased risk of chromosomal abnormalities. These characteristics are shared with the equality variable, with which symmetry was closely correlated, but only equality resulted significant in the multiple analysis.
It should be underlined that all visual real-time scoring procedures are affected by varying inherent difficulties, i.e. the intra- and inter-observer variations are likely to be larger for some variables than others. Such qualities in a parameter may diminish its prognostic power, even if the variable is of significant biological importance. Presumably symmetry is the least precise variable in this sense. On the other extreme among the tested variables is cleavage rate. The low risk of mis-scoring cleavage rate is probably one of the factors that put this variable into the best scoring position. Thus, it should be kept in mind that the scoring model derived from the present study should be regarded primarily as a clinical tool for selecting embryos to transfer. The biological significance of the findings must be supported and tested in further studies.
The embryo score that was derived from the regression analysis was validated on the group with a single implantation, i.e. the treatments in which the exact fate of the individual embryo was not traceable, as one implantation resulted from two transferred embryos. The mean embryo score in that group was in between the estimated levels for the groups with completely failed implantation and twin implantation, respectively. As expected, also, the mean score for the best embryo (i.e. the embryo most likely to have implanted) in Group 1 was in the range of the scores for embryos in the twin group (with proven implantation), whereas the mean score for the lowest scoring embryo (most likely the embryo that generally did not implant) in Group 1 was in the range of the embryos in the group with completely failed implantation. The most likely scenario predicted for group 1 was thus observed, lending further support to the validity of the scoring model. Another indication of the model’s validity is the strong correlation found between IMC scores and the observed implantation rates in Groups 0 and 2. Naturally, a more powerful validation of the IMC scores is to perform independent prospective validation studies. This has subsequently been performed in a study comprising 1950 treatments in the clinic, showing excellent congruence with these results (Holte, 2006).
Age did not interact with any of the variables, i.e. the model can be applied to any patient independent of age. An attempt to fit the age-specific results to an algorithm (Figure 3) showed that implantation figures decrease proportionally with age, similarly at all embryo scores. However, for the purpose of a full clinical prediction model for implantation chance, the embryo scoring model must be combined with other relevant clinical data, similarly tested multivariately, in analogy with the design of other studies in this field (Roseboom et al., 1995; Terriou et al., 2001; Hunault et al., 2002; Peterson et al., 2004; Thurin et al., 2005). When constructing such a model, we found the embryo score to qualify as one of the few independent variables among many univariately significant variables. Such variables were, apart from age and IMC score, a measurement of ovarian sensitivity to FSH (reflecting ovarian age) and the number of previous successful/failed treatments. Prospective application of that model to select for single- or double-embryo transfer showed the model to be highly efficient as a means to reduce twin implantation rates (from 28 to 2%) at preserved pregnancy rates (Holte et al., 2004). A subsequently performed multicentre validation study, comprising another 
4000 treatments, further supports the power of that model (Holte, 2006).
In conclusion, the findings of this large prospective study of the ability of five real-time visual embryo scoring variables to predict implantation potential in the individual embryo suggest that cleavage stage (in a non-linear manner), scoring for equality of blastomere size and the percentage of mononucleated blastomeres (both linearly) have independent prognostic power. It is suggested that applying this IMC score model incorporating the three independent variables, as presented here, should form the basis for ranking and selecting embryos for transfer and thus optimize implantation figures.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
We thank Felicia Franzén for help with the statistics. Furthermore, we thank Anna Blomgren, Anne Picki, Ola Sundström, Marina Lilja, Ulrika Sandkvist, Janet Candel, Birgitta Eriksson and Ulla Bäckros, all BMAs, for their skilful laboratory work throughout this study.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Alikani M, Cohen J, Tomkin G, Garrisi GJ, Mack C, Scott RT. (1999) Human embryo fragmentation in vitro and its implications for pregnancy and implantation. Fertil Steril 71:836–842.[CrossRef][Web of Science][Medline]
Bergh T, Ericson A, Hillensjo T, Nygren KG, Wennerholm UB. (1999) Deliveries and children born after in-vitro fertilisation in Sweden 1982–95: a retrospective cohort study. Lancet 354:1579–1585.[CrossRef][Web of Science][Medline]
Child TJ, Abdul-Jalil AK, Tan SL. (2002) Embryo morphology, cumulative embryo score, and outcome in an oocyte in vitro maturation program. Fertil Steril 77:424–425.[CrossRef][Web of Science][Medline]
Claman P, Armant DR, Seibel MM, Wang TA, Oskowitz SP, Taymor ML. (1987) The impact of embryo quality and quantity on implantation and the establishment of viable pregnancies. J In Vitro Fert Embryo Transf 4:218–222.[CrossRef][Web of Science][Medline]
Copperman AB, Selick CE, Grunfeld L, Sandler B, Bustillo M. (1995) Cumulative number and morphological score of embryos resulting in success: realistic expectations from in vitro fertilization-embryo transfer. Fertil Steril 64:88–92.[Web of Science][Medline]
Cummins JM, Breen TM, Harrison KL, Shaw JM, Wilson LM, Hennessey JF. (1986) A formula for scoring human embryo growth rates in in vitro fertilization: its value in predicting pregnancy and in comparison with visual estimates of embryo quality. J In Vitro Fert Embryo Transf 3:284–295.[CrossRef][Medline]
Dean NL, Phillips SJ, Buckett WM, Biljan MM, Tan SL. (2000) Impact of reducing the number of embryos transferred from three to two in women under the age of 35 who produced three or more high-quality embryos. Fertil Steril 74:820–823.[CrossRef][Web of Science][Medline]
Dor J, Rudak E, Mashiach S, Nebel L, Serr DM, Goldman B. (1986) Periovulatory 17 beta-estradiol changes and embryo morphologic features in conception and nonconceptional cycles after human in vitro fertilization. Fertil Steril 45:63–68.[Web of Science][Medline]
Ebner T, Yaman C, Moser M, Sommergruber M, Polz W, Tews G. (2001) Embryo fragmentation in vitro and its impact on treatment and pregnancy outcome. Fertil Steril 76:281–285.[CrossRef][Web of Science][Medline]
Ebner T, Moser M, Sommergruber M, Tews G. (2003) Selection based on morphological assessment of oocytes and embryos at different stages of preimplantation development: a review. Hum Reprod Update 9:251–262.
Fisch JD, Rodriguez H, Ross R, Overby G, Sher G. (2001) The Graduated Embryo Score (GES) predicts blastocyst formation and pregnancy rate from cleavage-stage embryos. Hum Reprod 16:1970–1975.
Gabrielsen A, Lindenberg S, Petersen K. (2001) The impact of the zona pellucida thickness variation of human embryos on pregnancy outcome in relation to suboptimal embryo development. A prospective randomized controlled study. Hum Reprod 16:2166–2170.
Garside WT, Loret de Mola JR, Bucci JA, Tureck RW, Heyner S. (1997) Sequential analysis of zona thickness during in vitro culture of human zygotes: correlation with embryo quality, age, and implantation. Mol Reprod Dev 47:99–104.[CrossRef][Web of Science][Medline]
Giorgetti C, Terriou P, Auquier P, Hans E, Spach JL, Salzmann J, Roulier R. (1995) Embryo score to predict implantation after in-vitro fertilization: based on 957 single embryo transfers. Hum Reprod 10:2427–2431.
Hardarson T, Hanson C, Sjogren A, Lundin K. (2001) Human embryos with unevenly sized blastomeres have lower pregnancy and implantation rates: indications for aneuploidy and multinucleation. Hum Reprod 16:313–318.
Hill GA, Freeman M, Bastias MC, Rogers BJ, Herbert CM 3rd, Osteen KG, Wentz AC. (1989) The influence of oocyte maturity and embryo quality on pregnancy rate in a program for in vitro fertilization-embryo transfer. Fertil Steril 52:801–806.[Web of Science][Medline]
Holte J. (2006) A computer-based prediction model for eSET. ESHRE campus, Helsinki 2006: Elective single embryo transfer, why, when and how? Abstract.
Holte J, Bergh T, Tilly J, Pettersson H, Berglund L. (2004) The construction and application of a prediction model to minimize twin implantation rate at a preserved high pregnancy rate. Hum Reprod 19:Suppl 1, 2004: Abstract Book. 20th Annual Meeting of the ESHRE, P-394.
Hunault CC, Eijkemans MJ, Pieters MH, te Velde ER, Habbema JD, Fauser BC, Macklon NS. (2002) A prediction model for selecting patients undergoing in vitro fertilization for elective single embryo transfer. Fertil Steril 77:725–732.[CrossRef][Web of Science][Medline]
Laasch C and Puscheck E. (2004) Cumulative embryo score, not endometrial thickness, is best for pregnancy prediction in IVF. J Assist Reprod Genet 21:47–50.[CrossRef][Web of Science][Medline]
Magli MC, Gianaroli L, Ferraretti AP. (2001) Chromosomal abnormalities in embryos. Mol Cell Endocrinol 183:Suppl 1, S29–S34.
Moon SY, Kim SH, Jung BJ, Jee BC, Suh CS, Lee JY. (2000) Influence of female age on pregnancy outcome in in vitro fertilization and embryo transfer patients undergoing intracytoplasmic sperm injection. J Obstet Gynaecol Res 26:49–54.[Medline]
Palmstierna M, Murkes D, Csemiczky G, Andersson O, Wramsby H. (1998) Zona pellucida thickness variation and occurrence of visible mononucleated blastomers in preembryos are associated with a high pregnancy rate in IVF treatment. J Assist Reprod Genet 15:70–75.[Web of Science][Medline]
Peterson CM, Reading JC, Hatasaka HH, Parker Jones K, Udoff LC, Adashi EY, Kuneck PH, Erickson LD, Malo JW, Campbell BF, et al. (2004) Use of outcomes-based data in reducing high-order multiple pregnancies: the role of age, diagnosis, and embryo score. Fertil Steril 81:1534–1541.[CrossRef][Web of Science][Medline]
Puissant F, Van Rysselberge M, Barlow P, Deweze J, Leroy F. (1987) Embryo scoring as a prognostic tool in IVF treatment. Hum Reprod 2:705–708.
Roseboom TJ, Vermeiden JP, Schoute E, Lens JW, Schats R. (1995) The probability of pregnancy after embryo transfer is affected by the age of the patient, cause of infertility, number of embryos transferred and the average morphology score, as revealed by multiple logistic regression analysis. Hum Reprod 10:3035–3041.
Saldeen P and Sundstrom P. (2005) Nuclear status of four-cell preembryos predicts implantation potential in in vitro fertilization treatment cycles. Fertil Steril 84:584–589.[CrossRef][Web of Science][Medline]
Shulman A, Ben-Nun I, Ghetler Y, Kaneti H, Shilon M, Beyth Y. (1993) Relationship between embryo morphology and implantation rate after in vitro fertilization treatment in conception cycles. Fertil Steril 60:123–126.[Web of Science][Medline]
Staessen C, Camus M, Bollen N, Devroey P, Van Steirteghem AC. (1992) The relationship between embryo quality and the occurrence of multiple pregnancies. Fertil Steril 57:626–630.[Web of Science][Medline]
Steer CV, Mills CL, Tan SL, Campbell S, Edwards RG. (1992) The cumulative embryo score: a predictive embryo scoring technique to select the optimal number of embryos to transfer in an in-vitro fertilization and embryo transfer programme. Hum Reprod 7:117–119.
Stromberg B, Dahlquist G, Ericson A, Finnstrom O, Koster M, Stjernqvist K. (2002) Neurological sequelae in children born after in-vitro fertilisation: a population-based study. Lancet 359:461–465.[CrossRef][Web of Science][Medline]
Terriou P, Sapin C, Giorgetti C, Hans E, Spach JL, Roulier R. (2001) Embryo score is a better predictor of pregnancy than the number of transferred embryos or female age. Fertil Steril 75:525–531.[CrossRef][Web of Science][Medline]
Thurin A, Hardarson T, Hausken J, Joblonowska B, Lundin K, Pinborg A, Bergh C. (2005) Predictors of ongoing implantation in IVF in a good prognosis group of patients. Hum Reprod 20:1876–1880.
Van Blerkom J, Davis P, Alexander S. (2001) A microscopic and biochemical study of fragmentation phenotypes in stage-appropriate human embryos. Hum Reprod 16:719–729.
Van Royen E, Mangelschots K, De Neubourg D, Valkenburg M, Van de Meerssche M, Ryckaert G, Eestermans W, Gerris J. (1999) Characterization of a top quality embryo, a step towards single-embryo transfer. Hum Reprod 14:2345–2349.
Van Royen E, Mangelschots K, De Neubourg D, Laureys I, Ryckaert G, Gerris J. (2001) Calculating the implantation potential of day 3 embryos in women younger than 38 years of age: a new model. Hum Reprod 16:326–332.
Van Royen E, Mangelschots K, Vercruyssen M, De Neubourg D, Valkenburg M, Ryckaert G, Gerris J. (2003) Multinucleation in cleavage stage embryos. Hum Reprod 18:1062–1069.
Visser DS and Fourie FR. (1993) The applicability of the cumulative embryo score system for embryo selection and quality control in an in-vitro fertilization/embryo transfer programme. Hum Reprod 8:1719–1722.
Ziebe S, Petersen K, Lindenberg S, Andersen AG, Gabrielsen A, Andersen AN. (1997) Embryo morphology or cleavage stage: how to select the best embryos for transfer after in-vitro fertilization. Hum Reprod 12:1545–1549.
Ziebe S, Lundin K, Loft A, Bergh C, Nyboe Andersen A, Selleskog U, Nielsen D, Grondahl C, Kim H, Arce JC. (2003) FISH analysis for chromosomes 13, 16, 18, 21, 22, X and Y in all blastomeres of IVF pre-embryos from 144 randomly selected donated human oocytes and impact on pre-embryo morphology. Hum Reprod 18:2575–2581.
Submitted on February 27, 2006; resubmitted on September 6, 2006; accepted on September 12, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Brodin, T. Bergh, L. Berglund, N. Hadziosmanovic, and J. Holte High basal LH levels in combination with low basal FSH levels are associated with high success rates at assisted reproduction Hum. Reprod., November 1, 2009; 24(11): 2755 - 2759. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Racowsky, L. Ohno-Machado, J. Kim, and J. D. Biggers Is there an advantage in scoring early embryos on more than one day? Hum. Reprod., September 1, 2009; 24(9): 2104 - 2113. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Devroey, B.C.J.M. Fauser, K. Diedrich, and on behalf of the Evian Annual Reproduction (EVAR) Approaches to improve the diagnosis and management of infertility Hum. Reprod. Update, July 1, 2009; 15(4): 391 - 408. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A.M. van Peperstraten, W.L.D.M. Nelen, R.P.M.G Hermens, L. Jansen, E. Scheenjes, D.D.M. Braat, R.P.T.M. Grol, and J.A.M. Kremer Why don't we perform elective single embryo transfer? A qualitative study among IVF patients and professionals Hum. Reprod., September 1, 2008; 23(9): 2036 - 2042. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R.W. Browne, W.B. Shelly, M.S. Bloom, A.J. Ocque, J.R. Sandler, H.G. Huddleston, and V.Y. Fujimoto Distributions of high-density lipoprotein particle components in human follicular fluid and sera and their associations with embryo morphology parameters during IVF Hum. Reprod., August 1, 2008; 23(8): 1884 - 1894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P.O. Karlstrom and C. Bergh Reducing the number of embryos transferred in Sweden-impact on delivery and multiple birth rates Hum. Reprod., August 1, 2007; 22(8): 2202 - 2207. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||