Embryo fragmentation and morphokinetics
By using the cell division timing grouped quartiles (previously published by our group and presented in the Supplemental Table S1), we found that fragmentation and morphokinetics were independent variables for clinical pregnancy in embryos which had up to 32% final fragmentation. The different quartiles represent different implantation potential. Embryos with t2 and with tSB in quartile 1 (Q1) had a higher chance to implant than embryos with t2 and tSB in quartile 4 [24]. The t2 and tSB of most of the fragmented embryos were equally distributed in the first 3 quartiles (Table 1A) and were not concentrated in quartile 3 or 4 as one may assume.
The percentage of start and final fragmentation was similar in quartiles 1-3 for t2 (Table 1A). Embryos with a start fragmentation high as 23.7±18.8 and final fragmentation high as 32.5±21.2 were significantly more prevalent in quartile 4 (Table 1A). About 65% of embryos from the first three quartiles were transferred or frozen, while from the 4th quartile, only 47% were transferred or frozen (p<0.003). Overall, these embryos cleaved slower into two cells followed by the appearance of more than 50% fragmentation.
For the tSB timing, no difference was found between start fragmentation, final fragmentation and fragmentation worsening in all four quartiles of timing. One hundred and sixty-five embryos were defined as Stage 0 - embryos that did not reach the start blastulation (SB) developmental stage and therefore had no tSB timing. These embryos had a significantly higher start and final fragmentation compared with embryos that reached the SB stage (at Q1-4) (p<0.0001) (Table 1B).
Fragmentation and fragmentation worsening
No correlation was found between fragmentation worsening and clinical pregnancy and live birth. However, we found that 1.76 was the cutoff value of fragmentation worsening for embryo transfer or freezing..
The average fragmentation following first cell division (the start fragmentation) in all 379 embryos was 17.9% ±14.8% and the final (approximately 110 hours post insemination) was 25.5% ±18.0%. Those embryos were divided into 3 groups according to the start fragmentation. The 3 groups were set according to similar groupings previously reported [14, 17]. The first group (n= 227 embryos) consisted of embryos with up to 15% at the start of fragmentation, the second group (n=111 embryos) with 15%-35% fragmentation and the third group (n=41 embryos) with more than 35% fragmentation from the first cell division. Fragmentation worsening was calculated for each group. A significant difference was found in fragmentation worsening between embryos with start fragmentation of up to 15% and embryos with start fragmentation of 15%-35% (worsening of x1.9±1.3 and x1.4±0.4 respectively p<0.001). We found that above 35% fragmentation at first cell division there was no point in calculating fragmentation worsening since there was already a high degree of fragmentation to begin with.
Embryo Stage on Day 5 and Fragmentation
Embryos were divided according to their developmental stage at Day 5; group 0 for embryos that did not reach the morula stage, group 1 for embryos at the morula stage, group 2 for start blastocyst, group 3 for blastocyst and group 4 for expanded blastocyst stage (Table 2). Significant difference in percentage of fragmentation was found between embryos which did not develop to the morula stage on Day 5 (group 0) compared with embryos that continued to develop to morula, start blastocyst, blastocyst, or expanded blastocyst stages (groups 1-4) (p<0.0001). This group of embryos (group 0) had high fragmentation (28.9%±19.5% at start fragmentation and 42.7%±19.2% at their final fragmentation) compared with other embryos which developed at least to morula stage (groups 1-4) (Table 2).
In embryos which developed to the morula, start blastocyst, blastocyst, and expanded blastocyst stage at Day 5 (groups 1-4), a similar start fragmentation was measured (Table 2). The average final fragmentation rates in embryos which reached the morula or start blastocyst stage were similar (groups 1-2) (19.4% ± 10.3% and 20.1% ± 9.8% respectively) but was significantly different from embryos at the blastocyst and expanded blastocyst stage (groups 3-4) (16.4±10.4% and 14.2%±5.8% respectively (p<0.005) (Table 2).
Fragmentation and embryo fate
Embryos which were transferred, or frozen had a significantly lower start fragmentation (12.0% ±7.1%) and final fragmentation (15.7% ±8.6%) compared with discarded embryos. The discarded embryos had 25.8%±18.5% start and 38.8%±18.9% final fragmentation (p<0.0001) (Table 3).
The fragmentation worsening was also found to be significantly higher among discarded embryos (x1.5) than among transferred or frozen embryos (x1.3) (p<0.0001) (Table 3).
In a retrospective examination, using ROC analysis we found that 19.5% fragmentation on first cell division was the cut-off for embryo selection [AUC: 0.76 95%CI (0.71-0.82)]. More than 88% of the embryos which were transferred, or frozen had up to 19.5% start fragmentation. For final fragmentation, the cut-off was 27% fragmentation [AUC: 0.88 95%CI (0.85-0.92)]. More than 92% of the embryos which were transferred, or frozen had up to 27% final fragmentation (Table 4 A,B).
Patient's age and embryo fragmentation
The mean age of patients with Day 5 embryos with more than 5% fragmentation (n=379 embryos) was 32.7±5.4 years; similar to the entire study population of patients with Day 5 embryos during the study period [33.0±5.0 years old (n=4,210 embryos)]. Among these fragmented embryos, a significant difference in fragmentation was found between patients aged <35 compared with ≥35 years; start fragmentation was 16.5%±13.4% and 20.8%±17.1% respectively (p<0.001) and final fragmentation 23.3%±17% and 30.2%±19.3% respectively (p<0.0001) (Table 5). A correlation test between patient age and fragmentation showed a low but statistically significant correlation (r=0.23; p<0.0001).
ICSI and IVF
One hundred and forty-six embryos (52 patients) were IVF embryos, and 233 (89 patients) were ICSI embryos. No difference was found in fragmentation between embryos from IVF or ICSI. The start fragmentation in IVF embryos was 19.6% ± 17.7% compared with 18.4% ± 13.8% in ICSI embryos. The final fragmentation was 27.1% ± 20.4% in IVF embryos compared to 26.8% ± 16.8% in ICSI embryos.
Pregnancy and live birth rates
One hundred forty-one cycles included fragmented embryos. From those cycles, as high as 129 cycles included fragmented embryos that was selected for transfer of for freezing (figure 1). Live birth (44%) and miscarriage (5%) rates were similar to the entire D5 embryo population (4,210 embryos).
Among the 379 fragmented embryos, 145 are KID embryos from which 30 embryos were KID positive resulted in a live birth (Figure 1).
Interestingly, we found that 83% of the patients who had fragmented embryos, had additional fragmented embryos in that cycle and 57% of those patients, had fragmented embryos in a previous cycle (for 31% of these patients, this was their first and only cycle).
KID positive embryos and fragmentation
No difference was found between start fragmentation, final fragmentation, and the fragmentation worsening, between the KID positive and KID negative embryos. All KID positive embryos which yield a live birth had up to 20% fragmentation except for one case with 43% final fragmentation (Figure 2).