Assessment of numeric abnormalities of X, Y, 18, and 16 chromosomes in preimplantation human embryos before transfer

Assessment of numeric abnormalities of X, Y, 18, and 16 chromosomes in preimplantation human embryos before transfer Santiago MunnC, PhD,B Khalid M. Sultan, MD; Heinz-Ulrich Weier, PhD: c James A. Grifo, MD, PhD,” Jadques Cohen, PhD,” and Zev Rosenwaks, MD” New York, New York, and San Francisco and Berkeley, California OBJECTIVE: Our purpose was to determine the feasibility of ascertaining aneuploidy for chromosomes X, Y, 18, and 16 by use of multiple-probe fluorescence in situ hybridization in blastomeres from preimplantation human embryos. STUDY DESIGN: A short fluorescence in situ hybridization procedure involving the simultaneous use of four deoxyribonucleic acid probes detected with red, green, blue, or a mixture of red and green fluorochromes was developed to determine numeric abnormalities of chromosomes X, Y, 18, and 16. Embryos underwent biopsy, and all or most cells were analyzed to distinguish true aneuploidy from mosaicism and to assess technique variations within the same embryo (n = 64). RESULTS: The analysis of all the blastomeres of an embryo was achieved in 91% of the embryos. Successful analyses including biopsy, fixation, and fluorescence in situ hybridization were achieved in 87.8% of the blastomeres. Of the four chromosomes tested, numeric aberrations were found in 23% and 42% of normally and abnormally developing embryos, respectively, including aneuploidy, polyploidy, haploidy, and mosaicism. When diploid embryos containing one or several tetraploid cells are counted as chromosomally abnormal, then 49% and 61% of normally and abnormally developing embryos, respectively, were chromosomally abnormal. Aneuploid embryos consisted of two monosomies for chromosome 16, one for chromosome 18, and a trisomy for chromosome 16. There was a tendency for aneuploidy to increase with maternal age. CONCLUSIONS: Fluorescence in situ hybridization is a more efficient method than cytogenetic analysis to study specific aneuploidies at preimplantation stages of development in human embryos. In addition, the preimplantation genetic diagnosis of two blastomeres per eight-cell embryo may be sufficient to ensure successful analysis of polyploidy, haploidy, and specific aneuploidies without endangering the survival of the embryo. The technique can, be easily modified to consider other chromosomes, including 13 and ?l . Because most chromosomally abnormal embryos do not develop to term, the use of this technique may increase the delivery rate per embryo by allowing only transfer of embryos normal for the tested chromosomes. Thie technique would be most useful for older women undergoing in vitro fertilization, because aneuploidy appears to increase with advancing maternal age. (AM J OBSTET GYNECOL 1995;172:1191-201.)

Key words: hybridization,

Monosomy, preimplantation in vitro fertilization

Chromosomal aberrations genesis and early embryonic

genetic

diagnosis,

occurring during gametodevelopment play a signif-

From The Center for Reproductive Medicine and Infertility The New York Hosbital-Cornell Medical Center.” the Division of Molecular Cytometqt Department of Laboratory hedicine, School if Medicine, University of California, San Francisco,’ and the Division of Lz;fe Sciences, Lawrence Berkeley Laboratory, University of California.” Presented by invitation at the Thirteenth Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 8-10, 1994. Reprint requests: Zev Rosenwaks, MD, The Center for Reproductive Medicine and Infertility, The New York Hospital-Corriell Medical Center, 505 E. 70th St., HT300, New York, NY 10021. Copyright 0 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/6/62682

mosaicism,

fluorescence

in situ

icant role in fetal 10~s.~ For example, in a recent survey’ the rate of chromosomal abnormalities in spontaneous abortions after in vitro fertilization (IVF) was reported to be approximately 60%. It is estimated that only 30% of human embryos during natural cycle conception are implanted and reach viability; only 15% of these losses are recognized miscarriages, whereas the remainder are lost at earlier stages.3 Wilcox et aL4 have estimated that 22% of embryos in natural-cycle pregnancies are lost before the third week of pregnancy. The observation that 37% of karyotyped preimplantation embryos after IVF are abnormal5 and the fact that most arrested embryos’ and morphologically abnormal embryos’ are chromosomally abnormal suggest 1191

1192

Mum6

April 1995 Am J Obstet Gynecol

et al.

that early fetal loss is mostly a result of chromosomal abnormalities. The fact that 4.7% of embryos reaching the seventh to eighth week of pregnancy are chromosomally abnormal’ indicates that most abnormal embryos are lost before the pregnancy is clinically recognized. The relevance of early embryo loss is particularly underscored by the observed implantation efficiency after IVF. Not only is the implantation rate per embryo relatively low (15% to 20% in the best units), but the implantation efficiency also diminishes with maternal age. Considering that the older ( > 35 years old) population has the highest incidence of trisomic deliverie?, ‘y it would be beneficial to assess the chromosomal makeup of preimplantation embryos before embryo transfer. A fast method for aneuploidy analysis of human preimplantation embryos by fluorescence in situ hybridization of single blastomeres could be used before replacement in the uterus. This type of preimplantation analysis would prevent the birth of aneuploid offspring. The elimination of chromosomal abnormalities in transferred embryos could therefore increase the implantation rate of older women, because data from oocyte donation suggests that the primary factor for declining fertility in older females is oocyte senescence.” Furthermore, the localization of specific chromosomal segments may allow the detection of chromosomal translocations in embryonic cells when recurrent pregnancy loss is the result of a known balanced translocation in one of the parents. This strategy would require previous identification of the involved chromosomal segments in parental cells and assessment of derivative chromosomes in the polar body and in the preimplantation embryo. Fluorescence in situ hybridization analysis of multiple chromosomes requires the simultaneous use of multiple deoxyribonucleic acid (DNA) probes labeled with different fluorochromes. Multiple-color fluorescence in situ hybridization has been performed in various tissues’2-‘4 and has been used for gender determination in preimplantation human embryos’5-‘s and in preimplantation diagnosis of aneuploidy, including gonosomes and 18, 13, and 21 chromosomes.‘g In this latter method two shades of red fluorochromes and a green fluorochrome were used alone or in combination to detect four different chromosome-specific probes in arrested human embryos. Nevertheless, that method was unable to differentiate overlapping signals stained in different shades of red. The purpose of this study was to introduce an extra fluorochrome color (blue) to solve that problem. A probe for chromosome 16 was applied, instead of the prior 1312 1, to assess the rate of chromosome 16 aneuploidy, because this chromosome con-

tributes to 30% of trisomies detected abnormal spontaneous abortions.”

in chromosomally

Material and methods Embryos. Embryos were obtained from patients undergoing in vitro fertilization at The Center for Reproductive Medicine and Infertility, Cornell University Medical College. The number of pronuclei per zygote was scored after insemination, and those with two pronuclei (monospermic) were allowed to cleave for 3 days. Normally developing embryos were those that reached the six-to-eight-cell stage by day 3 of development, with 5 15% fragmentation, even cells, and few or no vacuoles, granulation, or multinucleated blastomeres.” Morphologically normal embryos for this study came from two sources. The first consisted of embryos from patients undergoing preimplantation genetic diagnosis for X-linked genetic diseases. The gender of these embryos was determined by single-cell biopsy followed by fluorescence in situ hybridization. The embryos with the gender at risk (i.e., male) of being affected were not transferred and were used for this study. The second source of normally developing embryos were those donated for research by patients 240 years old. In these patients the four embryos of best morphologic and developmental characteristics were replaced, and any others were discarded or donated for research, because freezing of these embryos rarely results in pregnancy. These embryos may be investigated by fluorescence in situ hybridization according to a protocol approved by the Human Investigation Committee of The New York Hospital--Cornell Medical Center. In addition, arrested or morphologically abnormal embryos were also analyzed. The 42 patients from whom the embryos were obtained had consented to these procedures. The mean maternal age for these embryos was 36.7 t 3.8 years (range 28 to 43 years). Embryo biopsy and blastomere preparation. Biopsy was performed on embryos on day 4 after retrieval. A hole was drilledZ2 through the zona pellucida with acidified Tyrode’s solution (pH 2.4), and several blastomeres were removed from each embryo by micromanipulation.e3, p4 Biopsy was performed on all the cells of each embryo one by one. Blastomeres were fixed individually on glass slides by means of a slight modification of Tarkowski’s technique.17, 25 Each blastomere was placed in a culture dish containing hypotonic solution (1% sodium citrate in water, 6 mg/ml bovine serum albumin, Sigma Chemical Co., St. Louis) for 5 minutes at 25” C and then transferred into a small volume of hypotonic solution on the slide. Ten milliliters of fixative (acetic acid/methanol 1: 3) was dropped on top while the cell was observed under a stereomicroscope.

Volume 172, Number Am J Obstet Gynecol

4, Part 1

Mum6

et al.

1193

The fixative was spread by continuous and gentle blowing until the cytoplasm dissolved. The position of the nucleus was marked with a tungsten-carbide-tip pencil. Slides were dehydrated (70%, 85%, 95% ethanol, 2 minutes each) and kept at -20” C before fluorescence in situ hybridization. Fluorescence in situ hybridization. A modification of a previously described protocol was used.” DNA probes for chromosomes X, Y, and 18 labeled with fluorescent haptens (Imagenetics, Naperville, Ill.) were applied in combination with a biotin-labeled probe specific for chromosome 16 (Oncor, Gaithersburg, Md.). The hybridization targets for the 18, X, 16, and Y chromosome probes were a-satellite repeat clusters in the centromeric region and satellite III DNA at Yqh, respectively. The probes for chromosomes Y and 18 were labeled with red (CEPr Spectrum orange, Imagenetics) and green (CEPT Spectrum green, Imagenetics) fluorescent haptens, respectively. For chromosome X a 1: 1 mixture of probes labeled with red and green fluorescent haptens was used to obtain a yellow appearance through a dual-wavelength filter (ChromaTechnology, Brattleboro, Vt.) and a white appearance with a triple-wavelength filter (Imagenetics). The probe for chromosome 16 was labeled with biotin (Oncor) and showed by using methylcoumarine acetic acid (AMCA)labeled avidin D (Vector Laboratories, Burlingame, Calif.). AMCA has blue fluorescence. The hybridization solution consisted of 4.5 ml of 16, 1.3 ml of Y, 1.3 ml of 18, and 1.3 ml of X mix-specific probe solutions, added to 8 ml of hybridization mix II solution (component of the Spectrum CEPr system, Imagenetics).26 The CEPr probes already contained blocking DNA. The resulting 16.4 ml of hybridization solution was applied to a glass slide containing fixed blastomeres and covered with an 18 x 18 mm coverslip. It was placed on a slide warmer at 80” C for 8 minutes, sealed with rubber cement, and placed to hybridize at 37” C in a dark moist chamber for 4 to 6 hours. The coverslip was removed and the slide was immersed in 50% formamide, 2 x saline-sodium citrate buffer, pH 7.0, at 42” C for 15 minutes, followed by a wash of 10 minutes in 2 x salinel-sodium citrate buffer at 42” C and in 1 x 2-phenyl-5-(4-biphenylyl)1,3,4-oxadiazole solution (Oncor) at room temperature for 2 minutes. Next, 15 ml of AMCA-conjugated avidin D (8 mg/ml, Vector Laboratories) was applied to the slide, covered with a plastic coverslip, and incubated for 30 minutes at 37” C. The slide was then washed twice in 1 x 2-phenyl-5-(4-biphenylyl)-1,3,4-oxadiazole solution during P-minute intervals. Fifteen microliters of biotinylated anti-avidin D goat antibody (20 mg/ml, Vector Laboratories) was applied to the slide and covered with a plastic coverslip. After 30 minutes at 37” C the slide

Fig. 1, A to C. In situ hybridization of fluorochrome-labeled X-, Y-, 18-, and 16-chromosome-specific DNA probes. Chromosome X-specific probe was labeled with 1: 1 red/green fluorochrome mixture, chromosome 18 in green, chromosome Y in red, and probe specific for chromosome 16 in blue. A and B, Female blastomere trisomic for chromosome 16 and normal for gonosomes and chromosome 18. A was taken with doublepass filter showing only hybridization signals labeled in red or green and consequently shows two X-chromosome (orange to red)- and two 18-chromosome (yellow)-specific signals. B, Same blastomere through filter showing only hybridization signals labeled in green or blue and consequently two Xchromosome (green)-, two l&chromosome (also green)-, and three 16-chromosome (blue)-specific signals. C, Normal male blastomere displaying an X-chromosome (orange)-, a Y-chromosome (red)-, two 18-chromosome (green)-, and two 16-chromosome (blue)-specific signals *counterstained with chromomycin As.

1194

Munn6

April 1995 Am J Obstet Gym01

et al.

Table I. Fluorescence in situ hybridization efficiency in blastomeres Cells from Broken

64 embryos

(n = 555)

during biopsy (n = 11, 2.0%) Lost during fixation (n = 9, 1.6%)

Cells used for other Cells fixed (n = 498)

studies

(n = 37)

Without nucleus (n = 40) With nucleus (n = 458) Cells analyzed by fluorescence (n = 458) Cells damaged or without Cells with signals (n = 451) Cells that were fluorescence failures (n = 32, 7.1%)

in situ hybridization signal (n = 7, 1.5%) in situ hvbridization

False negative (n = 30) XOlSi81616 (n = 2) XX1801616 (n = 6) x01801616 in = X01818160 (n = 1) XXI818160 (n = 20) False positive (n = 2) XX1818181616 (n = 2) Cells with correct number of signals

ij

(n = 419)

was washed twice in 1 x 2-phenyl-5-(4-biphenylyl)1,3,4-oxadiazole solution for 2 minutes. Fifteen microliters of AMCA-conjugated avidin D (8 mg/ml, Vector Laboratories) was added and covered with a plastic coverslip. After 30 minutes of incubation the coverslip was removed and the slide was washed twice in 1 x 2-phenyl-5-(4-biphenylyl)-1,3,4-oxadiazole solution for 2 minutes. To counterstain the nuclei, 15 ml of chromomycin As (0.05 mg/ml, Sigma) in antifade solution*’ was added. Fluorescence microscopy was performed with a Nikon Optiphot microscope (Nikon, Melville, N.Y.). A dual filter set for simultaneous observation of fluorescein isothiocyanate and Texas red (ChromaTechnology) was used to visualize Spectrum orange and Spectrum green signals. With this filter the Y-chromosome-specific probes appeared in red, the 18 chromosome in green, and the X chromosome in white to yellow. In addition, a Nikon filter for 4’,6-diamidino-2phenylindole/Hoechst was used for AMCA observation, and a Nikon filter to visualize fluorescein isothiocyanate was used to detect chromomycin and to localize the cells (see Fig.1, A to C). Digital computer imaging was not used during the analysis. Scoring criteria and fluorescence in situ hybridization failure criteria. As suggested by Hopman et al.,” minor hybridization spots that had much lower fluorescence intensity were not scored, and spots found in close vicinity to one another, interconnected or in paired arrangements, were counted as one signal. The minor hybridization spots were considered to be crosshybridization signals to nontarget chromosomes, whereas paired arrangements were most probably sister chromatids or split signals. Blastomeres with a clear nucleus detected at phase

contrast and not found or covered by debris after fluorescence in situ hybridization were considered failures. Blastomeres with less (false negative) or more (false positive) than two specific signals per chromosome pair analyzed were considered fluorescence in situ hybridization errors unless one of the following cases applied. (1) Sibling blastomeres had extra or missing signals compensating for the ones missing or extra in,a given blastomere. We considered this embryo a compensated mosaic embryo. (2) Two or more blastomeres with missing signals in the case that the sum of all the signals from these blastomeres was double or a multiple of a full haploid complement were considered mosaic embryos in which one cell had split in two or more without prior DNA replication. (3) Embryos in which all the blastomeres had the same abnormality, for example aneuploid, haploid, or polyploid embryos were considered to be errors. (4) Blastomeres that had only a single signal per chromosome were considered haploid cells. (5) Blastomeres that had three or more signals per chromosome pair were considered polyploid cells. (6) Cells with cross-hybridization signals or split signals, as stated above for the scoring criteria, were not counted as extra signals. The same prior criteria were also used for polynucleated blastomeres, counting as a unity all the nuclei present in the same blastomere. Results

Blastomere biopsy and fixation. Biopsy was performed on 555 blastomeres from 64 human embryos. The percentage of blastomeres lost after biopsy and fixation were 2.0% (11/555) and 1.7% (g/544), respectively. With the exception of 40 anuclear blastomeres and 37 blastomeres used for other studies, the rest (n = 458) were analyzed by fluorescence in situ hybridization (Table I). Lymphocyte fluorescence in situ hybridization failure. Lymphocytes (n = 200) from a healthy 34-year-old woman were scored. All cells, with the exception of 30 (15%), showed an XX18181616 karyotype. Twelve (6%) were fluorescence in situ hybridization failures of chromosome 16, including 10 xX1818160, one that did not show any 16 chromosome, and one XX 18 18 16 16 16; 11 (5.5%) were failures of chromosome 18, including eight XX1801616, two XX1818181616, and one X0 180 16 16; and 8 (4%) were failures of chromosome X, including six X018181616, one XXXX18181616, and one X01801616 (this one counted in both X and 18 fluorescence in situ hybridization failure). Blastomere fluorescence in situ hybridization failure. After fluorescence in situ hybridization analysis of 458 blastomeres, 451 cells (98.5%) showed hybridization signals, and seven blastomeres were either damaged (n = 5) or without clear signals (n = 2). Of the cells with hybridization signals, 32 were fluorescence in

Volume 172, Number Am J Obstet Gynecol

Table

II.

4, Part 1

Summary

developing

Munrk

of X-,

monospermic

Y-, human

18-,

and

Mosaic embryos,

III.

No. Diploid 1 2 3 4 5 6 7 8 Polyploid 9

abnormalities

found

in normally

and

1195

abnormally

embryos

Morphologically normal Chromosomally normal Normal with 4N cells (mosaic Chromosomally abnormal Monosomy 160 Monosomy 160 and mosaic Monosomy 180 and mosaic Other diploid mosaics Polyploid 3N/6N mosaic Polyploid complex mosaic Morphologically abnormal Chromosomally normal Normal with 4N cells (mosaic Chromosomally abnormal Trisomy 16 Other diploid mosaics Polyploid no mosaics Polyploid 3N/6N mosaic Polyploid complex mosaic Haploid mosaic (NIBN)

Table

16-chromosome

et al.

not including

31 16 8 7 1 1 1 2 1 1 33 13 6 14 1 4 4 1 3 1

2N/4N)

2N/4N)

normal

embryos

with tetraploid

Analyzed

(51.6%) (25.8%) (22.6%) (embryo (embryo (embryos

1, Table III) 2, Table III) 3, 4, Table

(embryo

5, Table

III)

III)

(39.4%) (18.2%) (42.4%) (embryos

6 to 9, Table

(embryos

10 to

12, Table

III)

III)

cells

Karyotypes

mosaics 717 10113 717 717 818 616 llill 717

X01818160 (4)*, XXX1818160 (2), XXXl818160t XY1801616 (6), XYl80161616 (4) XX18181616 (4), X018181616, XXX18181616, XXl818160:: XX18181616 (5), XX 5§d[18] 1616, XXl8181616]] XY18181616]] (4), XXYY 4[18] 4[16]1, XXYY 3[18] 4[16]$. XYY1818 3[16], XX18181616 (3) xX18181. 6X 4[18]:. 1818 XYY 3[18] 0, XX180160, Xx18181616, YOl80 4[16] (3), XX1818 0 (2), 3Y 3[18] 1616 (3) XYl8181616, XY 3[18] 1616, XYlSO$. X0160 (2), YO1818160 (2)

XY 3[18]

1616

mosaics lo/lo

10 11

414

12

516

616

3X 3[18] 3[16] (3), 3X 3[18] 3[16]t, X01801616 (2)t, 180160, 3X1801616?, 3X18181616, XX 3[18] 3[16]t 7X 7[18] 7[16] (2), 4X 8[18] 3[16], 8X 16[16] 4[16] 4X4Y 10[18] 6[16]t, XY 4[18] 3[16], XXYY 4[18] 1616, XX3Y1801616, XXY1818 3[16], XXYYY 5[18] 4[16] X018181616 (2), 3X 4[18] 1616, 3X 3[18] 1616, 8X 8[18] 6[16]t

*Blastomere number. tTota1 number of chromosomes in polynucleated blastomeres. ZBlastomere considered fluorescence in situ hybridization error. §Total number of 18 or 16 chromosomes. ]]Dinucleated blastomere with two signals per chromosome pair

situ hybridization errors: two false negative for chromosome X, six false negative for chromosome 18, one false negative for chromosome X and 16, one false negative for chromosome X and 18, 20 false negative for chromosome 16, and two false positive for chromosome 18. The fluorescence in situ hybridization failure rate in this study was 7.1% (32145 1). When combined with the 1.5% (7/458) blastomeres not classified after fluorescence in situ hybridization, the failure rate was 8.6%. When embryo biopsy (2.0%) and blastomere fixation (1.6%) losses were included, the total percentage of blastomeres without reliable results was 12.2% (Table I).

per

nucleus.

Results were obtained in 100% of the embryos analyzed (n = 64). In addition, when embryos (n = 8) in which some cells were used for other studies are not included, 91% (5 l/56) of these embryos had all their cells analyzed. The other five embryos contained one or more of the seven cells that did not yield fluorescence in situ hybridization results. Embryos with chromosome abnormalities. Morphologically normal embryos (n = 31) and arrested or morphologically abnormal embryos (n = 33) were analyzed by fluorescence in situ hybridization. Embryos with chromosome abnormalities are shown in Table II.

1196

Mum&

April 1995 Am J Obstet Gynecol

et al.

Of the 31 morphologically normal embryos analyzed, 16 (5 1%) had a normal diploid number of gonosomes and chromosomes 18 and 16. Eight embryos showed normal diploid cells in combination with one to three tetraploid cells. For reasons discussed in the next section these embryos were also considered to be chromosomally normal. The remaining embryos (n = 7, 23%) showed numeric chromosomal aberrations (Table III). Three (10%) of these embryos were aneuploid, that is, all the cells from the same embryo had the same abnormal number of specific signals for a given chromosome pair. Also, two (numbers 1 and 2 in Table III) were mosaics for another chromosome pair. These embryos are further described below. Of the aneuploid embryos, one was monosomic for chromosome 18, and two were monosomic for chromosome 16. The remaining four abnormal embryos were mosaics, that is, embryos with two or more cell lines differing in the number of gonosomes, 18 or 16 chromosomes. Two of these four mosaics were mostly diploid (numbers 3 and 4 in Table III), and the other two were polyploid mosaics. Of the two polyploid mosaics, one was triploid with hexaploid cells, and the other was a complex mosaic (embryo 9, Table III). Morphologically abnormal embryos (n = 33) were normal in 39% (13/33) of the cases, whereas six embryos showed normal diploid cells in combination with one or more tetraploid cells. The remaining embryos (n = 14, 42%) were chromosomally abnormal. One seven-cell embryo (3%) was aneuploid with trisomy 16 in all seven cells. Four embryos that were mostly diploid mosaics are shown in Table III (embryos 5 to 8). Four were polyploid embryos: a three-cell triploid embryo (XXY 18 18 18 16 16 16), a five-cell tetraploid embryo (4X 4[ 181 4[ 16]), a one-cell tetraploid embryo with an extra 18 chromosome (4X 5[ 181 4[16]), and a one-cell embryo with an XXXYY 13[ 181 4[ 161 chromosome constitution. In addition, four other embryos were polyploid mosaics: a triploid/hexaploid mosaic and three complex mosaics shown in Table III (embryos 10 to 12). In addition, a five-cell embryo was haploid with two diploid cells. The mosaic embryos were probably generated by different mechanisms. Embryos 1 and 3 from Table III were probably produced by mitotic nondisjunction, because they contained monosomic and trisomic blastomeres in addition to normal cells. The mosaic origin of embryo two, which included disomic and trisomic cells for chromosome 16, is not clear, although it is unlikely that it resulted from fluorescence in situ hybridization failure, because 40% (4110) of these cells were trisomic. This percentage is much higher than the overall 5.7% fluorescence in situ hybridization failure for this study; embryo number four had five 18 chromosomes in one of its seven cells, which may indicate

endoreduplication of that chromosome instead of fluorescence in situ hybridization failure. Embryos six and eight could have been produced by an abnormal DNA replication followed by uneven karyokinesis, and the rest of the mosaic embryos could have been produced by any combination of the above mechanisms, including the occurrence of mosaicism in an already aneuploid embryo. The frequency of mosaicism would be higher if embryos with tetraploid cells were included. Altogether one haploid, one aneuploid, and 14 normal embryos (25%, 16/64) contained tetraploid cells representing 10% to 50% of their cells. A total of 13.3% of blastomeres (60/450) were multinucleated, with higher percentages in cells from abnormally developing embryos (19.3%, 331172) compared with normally developing embryos (9.7%, 271278) (x’, p < 0.01). These multinucleated blastomeres were found in 45% (15/33) and 39% (12131) of abnormally and normally developing embryos, respectively. In these embryos 10% to 50% of their cells were multinucleated blastocysts. Interestingly, 15% (S/60) of multinucleated blastocysts contained fluorescence in situ hybridization failures, compared with 5.9% (23/391) in mononucleated cells (x’, p < 0.025). The total aneuploidy rate for gonosomes 18 and 16 chromosomes was 6.25% with a tendency to increase with maternal age, although the difference in aneuploidy rates at different maternal ages was not significant because of the small sample size. These rates were 0% (O/16) in women 28 to 34 years old, 3.1% (1132, trisomy 16) in women 35 to 39 years old, and 18.8% (3/16, two monosomy 16 and one monosomy 18) in women >40 years old.

Comment Previous fluorescence in situ hybridization studies of human blastomeres have reported high efficiency rates and have demonstrated the suitability of this technique for the preimplantation diagnosis of X-linked diseases and aneuploidy. 15,“, 18,” However, the analysis of a single chromosome pair in those studies could not differentiate between trisomy and triploidy or between monosomy and haploidy. In this and an earlier study’. I9 the difficulty was overcome by simultaneously analyzing three or more chromosome pairs, with comparable efficiencies. Biopsy of two blastomeres from an eight-cell human embryo has been shown to be nondetrimenta13” Thus at 88% efficiency a two-blastomere biopsy should be adequate to detect numeric chromosome abnormalities involving gonosomes and 18 and 16 chromosomes. One of the exceptions is mosaic embryos; for proper analysis all or most of the cells should be tested, which is obviously incompatible with embryo survival. In this study mosaics were detected because most or all the cells from each embryo were indeed tested. Similarly, aneuploidy was

Volume 172, Number Am J Obstet Gynecol

4, Part 1

differentiated from mosaicism by obtaining results in all embryos and from all or most cells of each embryo. In contrast, classic cytogenetic analysis of preimplantation embryos produces results in a low number of embryos and in only a fraction of the blastomeres analyzed. For example, in one of the largest cytogenetic studies on preimplantation human embryos only 29% of the embryos produced results, and only 25% of the embryos with results could have all their cells analyzed.’ Furthermore, the typical excess of missing chromosomes attributed to the fixation of metaphase nuclei and experienced in most karyotype studies of oocytes and embryos7. 3’ frequently does not allow the differentiation of aneuploidy from mosaicism, much less determine mechanisms of mosaicism formation. Loss of DNA, which produces false negatives, also occurs when fluorescence in situ hybridization is used, but on a much smaller scale. Fluorescence in situ hybridization obviously has the disadvantage of supplying only information on the chromosomes for which specific probes were used, but for these chromosomes the information is more reliable. For example, the four aneuploid embryos detected in this study showed the same trisomy or monosomy in all their cells, with seven to eleven cells analyzed per embryo and only one embryo where not all the cells were analyzed (embryo 2, Table III). If embryos (n = 8) in which some cells were used for other studies are not included, only 9% (5156) of the embryos studied by fluorescence in situ hybridization did not have all their cells analyzed. The fact that more multinucleated blastocysts contained failures of fluorescence in situ hybridization may be caused by two different mechanisms. One, dinucleated blastomeres normally show smaller nuclear diameters, and consequently the chances of two signals of the same color overlapping may increase. Two, many multinucleated blastocysts contain micronuclei that can get lost during fixation, and consequently we would be overestimating the fluorescence in situ hybridization failure rate and underestimating the fixation failure rate. Mosaicism has been found in high rates in preimplantation embryos,7, S’s 33 and it was the most common chromosome abnormality found in this study, in both normally and abnormally developing embryos. In contrast to karyotype studies, the origin of the mosaicism mechanisms can be distinguished in many of these embryos, including mitotic nondisjunction (embryos 1 and 3 in Table III), zero to several DNA replications followed by random karyokinesis (embryos 5, 6, 8, and 9, Table III), and probably anaphase lag of a chromosome 16 in an embryo that was previously trisomic for 16 and monosomic for 18 chromosomes (embryo 2, Table III). But the origin of most polyploid mosaic embryos (embryos 10, 11, and 12) and embryos 4 and 7 (Table III) was not clarified. In addition, numerically

Munnb

et al.

1197

imbalanced mosaic embryos (n = 14) consisted of normal diploid embryos with 10% to 50% of tetraploid cells. The occurrence of tetraploid cells in normal diploid embryos from cleaving to blastocyst stage has been described before.6. 34 Although it appears to be a normal phenomenon in the trophectoderm of many mammalian blastocysts, including the human,34-3’ its relevance in cleaving embryos is not yet understood, and they may represent either precursor cells of the trophectoderm or karyokinesis failure. The importance of mosaicism in cleaving embryos may depend on the number of chromosomally abnormal cells. Mosaic embryos containing a sizable number of normal cells may be able to develop fairly well, either because abnormal cells have a proliferative disadvantage, such as monosomic cells,” or because a percentage of normal cells is enough to produce a normal phenotype. For example, artificially produced mouse chimeras have resulted in phenotypically normal fetuses containing a percentage of chromosomally abnormal cells in most tissues.“. 39 On the other hand, on the basis of totipotency evaluations from embryo biopsy and freeze-thaw data30. 4o, ‘l embryos that have lost three of eight cells are able to develop normally to term. In normally developing embryos 10 of 14 mosaics, including 2N/4N mosaics, had ~38% (three of eight cells) chromosomally abnormal cells compared with eight of 15 mosaics in abnormally developing embryos (this study). Consequently, although mosaicism was high in both groups, milder forms of mosaicism were found more often in normally than in abnormally developing embryos. Similarly, multinucleated blastocysts were also more common in morphologically abnormal embryos. The kinds of multinucleated blastocyst karyotypes found in this study corresponded to the same ones reported by us in a previous work4* and suggest that no other mechanisms of multinucleated blastocyst formation are present. Aneuploidy for gonosomes and chromosomes 16 and 18 has been found in 6.25% of the embryos analyzed, with three of the four aneuploid embryos being monosomic. With the exception of monosomy 21, with an incidence of one in 1000 karyotyped abortions,43 the rest of autosomal monosomies are practically undetected in clinically recognized pregnancies. By use of karyotype analysis or simultaneous fluorescence in situ hybridization analysis of multiple chromosome pairs, several specific monosomies have been described in preimplantation human embryos, with at least two cells analyzed per embryo showing the same result: monosomy X0,“. 33 Y0,6 monosomy 15,44 monosomy 18,19 and monosomy 13 or 2 1 .19To our knowledge, this is the first time that monosomy 16 has been described in human embryos. Because nondisjunction produces as many disomic as nullisomic gametes, monosomic embryos must be eliminated during the first days or weeks

1198

Munn~

et

al.

April

Am J Obstet

of pregnancy, as in the mouse.45 The scarce data on autosomal monosomy found in preimplantation embryos and the technical errors produced in the detection of nullisomic gametes does not yet allow further discussion on monosomies before implantation. Similarly, although the results indicate a tendency for aneuploidy to increase with maternal age, no significant differences were obtained. Consequently, it is still not known whether the increase of aneuploidy with maternal age is because of an increase in aneuploid gametes or a deficient selection of trisomic offspring during gestation, as the two most prominent hypotheses argue (Warburton 198946). The current results demonstrate the feasibility of a four-color fluorescence in situ hybridization technique to detect aneuploid human embryos. By using only two fluorochromes in a previous study, we were able to analyze X, Y, 18, and 1312 1 chromosomes simultaneously.” By combining these two techniques, human embryos could be assessed for the most common aneuploidies in a reasonable time frame compatible with regular IVF procedures. The application of fluorescence in situ hybridization with multiple probes in IVF may prevent the replacement of aneuploid embryos and other chromosomally abnormal embryos with limited chances of implantation. Because embryonic aneuploidy is believed to increase with maternal age’, lo older patients would derive the most benefit.

11.

12.

13.

14.

15.

16.

17.

18.

19. REFERENCES 1. Chard T. Frequency of implantation and early pregnancy loss in natural cycles. In: Sepal% M, ed. Factors of importance for implantation. London: Bailliere Tindall, 1991: 179-82. M. Chromosome analysis of spontaneous abor2. Plachot tions after IVF: a European survey. Hum Reprod 1989;5: 425-9. 3. Lidley M. Life and death before birth [Editorial]. Nature 1979;280:635. 4. Wilcox AJ, Weinberg CR, O’Connor JF, et al. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189-94. 5. Plachot M, Veiga A, Montagut J, et al. Are clinical and biological IVF parameters correlated with chromosomal disorders in early life: a multicentric study. Hum Reprod 1988;5:627-35. 6. MunnC S, Grifo J, Cohen J, Weier HUG. Chromosome human arrested preimplantation abnormalities in embryos: a multiple-probe FISH study. Am J Hum Genet 1994;55:150-9 7. Pellestor F, Dufour MC, Arnal F, Humeau C. Direct assessment of the rate of chromosomal abnormalities in grade IV human embryos produced by in-vitro fertilization procedure. Hum Reprod 1994;9:293-302. 8. Burgoyne PS, Holland K, Stephens R. Incidence of numerical chromosome abnormalities in human pregnancy estimated from induced and spontaneous abortion data. Hum Reprod 1991;6:555-65. 9. Hassold T, Chiu D. Maternal age-specific rates of numerical chromosome abnormalities with special reference to trisomy. Hum Genet 1985;70: 11-7. 10. Warburton D, Kline J, Stein 2, Strobino B. Cytogenetic

20.

21.

22.

23.

24

25.

26.

27.

28.

1995

Gynecol

abnormalities in spontaneous abortions of recognized conceptions. In: Porter IH, Willey A, eds. Perinatal genetics: diagnosis and treatment. New York: Academic, 1986: 13348. Sauer Mv, Paulson RJ, Lobo RA. Reversing the natural decline in human fertility: an extended clinical trial of oocyte donation to women of advanced reproductive age. JAMA 1992;268:1275-9. Nederlof PM, Robinson D, Abuknesha R, et al. Three color fluorescence in-situ hybridization for the simultaneous detection of multiple nucleic acid sequences. Cytometry 1989;10:20-7. Dauwerse JG, Wiegant J, Raap AK, Breuning MH, Van Ommen GJB, Multiple colors by fluorescence in-situ hybridization using ratio-labelled DNA probes create a molecular karyotype. Hum Mol Genet 1992;1:593-8. Ried T, Baldini A, Rand T, Ward DC. Simultaneous visualization of seven different DNA probes by in-situ hybridization using combinational fluorescence and digital imaging microscopy. Proc Nat1 Acad Sci U S A 1992;89: 1388-92. Griffin DK, Wilton LJ, Handyside AH, Wiston RML, Delhanty JDA. Dual fluorescent in-situ hybridization for simultaneous detection of X and Y chromosome-specific probes for the sexing of human preimplantation embryonic nuclei. Hum Genet 1992;89:18-22. Delhanty JDA, Griffin DK, Handyside AH, et al. Detection of aneuploidy and chromosomal mosaicism in human embryos during preimplantation sex determination by fluorescence in situ hybridization (FISH). Hum Mol Genet 1993;2:1183-5. Munne S, Weier HUG, Stein J, Grifo J, Cohen J. A fast and efficient method for simultaneous X and Y in-situ hybridization of human blastomeres. J Assist Reprod Genet 1993;10:82-90. Mu&C S, Tang YX, Grifo J, Rosenwaks Z, Cohen J. Sex determination of human embryos using the polymerase chain reaction and confirmation by fluorescence in situ hybridization. Fertil Steril 1994; 11 l-7. MunnC S, Lee A, Rosenwaks Z, Grifo J, Cohen J. Diagnosis of major chromosome aneuploidies in human preimplantation embryos. Hum Reprod 1993;8:2185-91. Hassold TJ, Pettay D, Freeman SB, Grantham M, Takaesu N. Molecular studies of nondisjunction in trisomy 16. J Med Genet 1991;28:159-62. Cohen J, Malter HE, Talansky BE, Grifo J. Micromanipulation of human gametes and embryos. New York: Raven, 1992:223-95. Gordon JW, Talansky BE. Assisted fertilization by zona drilling: a mouse model for correction of oligospermia. J Exp 2001 1986;239:347-54. Grifo JA, Tang YX, Munnt S, Alikani M, Cohen J, Rosenwaks Z. Healthy deliveries from biopsied human embryos. Hum Reprod 1994;9:912-6 . Grifo JA, Tang YX, Cohen J, Gilbert F, Sanyal MK, Rosenwaks 2. Ongoing pregnancy in a hemophilia carrier by embryo biopsy and simultaneous amplification of X and Y chromosome specific DNA from single blastomeres. JAMA 1992;6:727-9. Tarkowski AK. An air drying method for chromosome preparations from mouse eggs. Cytogenetics 1966;5:394400. Pinkel D, Gray JW, Trask B, van den Engh G, Fuscoe J, van Dekken H. Cytogenetic analysis by in-situ hybridization with fluorescently labeled nucleic acid probes. Cold Spring Harb Symp Quant Biol 1986;51:151-7. Tohnson GD, de C Nogueira Arauio GM. A simule method of reducing the fad&g of immunofluorescence during microscopy. J Immunol Methods 1981;43:349-50. Hopman AHN, Ramaekers FCS, Raap AK, et al. In-situ hybridization as a tool to study numerical chromosome aberrations in solid bladder tumors. Histochemistry 1988; 89:307-16. _I

I

Volume 172, Number Am J Obstet

4, Part 1

29. Schrurs BM, Winston RML. Handvside AH. Preimolantation diagnosis of aneupldidy using fluorescent in-situ hybridization: evaluation using a chromosome 18-specific probe. Hum Reprod 1993;8:296-301. 30. Hardy K, Martin KL, Leese HJ, Winston RML, Handyside AH. Human preimplantation development in-vitro is not adversely affected by biopsy at the 8-cell stage. Hum Reprod 1990;5:708-14. 31. Zenzes MT, Casper RF. Cytogenetics of human oocytes, zygotes, and embryos after in-vitro fertilization. Hum Genet 1992;88:367-75. 32. Plachot M, Junta AM, Mandelbaum J, De Grouchi J, Salat-Baroux J, Cohen J. Chromosome investigations in early life: human preimplantation embryos. Hum Reprod 1987;2:29-35. 33. Bongso A, Fong CH, Ng SC, Ratman S, Lim J. Preimplantation genetics: chromosomes of fragmented human embryos. Fertil Steril 1991;56:66-70. 34. Benkhalifa M, Janny L, Vye P, Malet P, Boucher D, Menezo Y. Assessment of ploidy in human morulae and blastocysts using co-culture and fluorescent in-situ hybridization. Hum Reprod 1993;8:895-902. 35. Murray JD, Moran C, Boland MP, et al. Polyploid cells in bIasto&& and early fetuses from Au&&an Merino sheep. I Reprod Fertil 1986:78:439-46. 36. Long Sk, W>lliams CV. A comparison of the chromosome complement of inner cell mass and trophoblast cells in day-i0 pig embryos. J Reprod Fertil 1982;66:645-8. 37. Hare WCD, Singh EL, Betteridge KJ, et al. Chromosomal analysis of 159 bovine embryos collected 12 to 18 days after es&us. Can J Genet Cytol 1980;22:615-26. 38 Epstein CJ, Smith S, Cox DR. Production and properties of mouse trisomy 15 diploid chimeras. Dev Genet 1984; 4:159-65. 39. Epstein CJ, Smith SA, Zamora T, Sawicki JA, Magnuson TR. Cox DR. Production of viable adult trisomv 17 dinloid mo&e chimeras. Proc Natl Acad Sci U S A 1962;79:437680. 40.

41.

42.

43.

44.

Munnb

et al.

1199

Gynecol

Wilton LJ, Shaw JM, Trounson AO. Successful single-cell biopsy and cryopreservation of preimplantation mouse embryos. Fertil Steril 1989;51:513-7. Veiga A, Calderon G, Barri PN, Coroleu B. Pregnancy after the replacement of a frozen-thawed embryo with < 50% intact blastomeres. Hum Reprod 1987;2:321-7. Munnt S, Cohen J. Unsuitability of multinucleated human blastomeres for preimplantation genetic diagnosis. Hum Reprod 1993;8:1120-5. Hassold T, Chen N, Funkhouser J, et al. A cytogenetic study of 1000 spontaneous abortuses. Ann Hum Genet 1980;44:151-78. Angel1 RR, Aitken RJ, van Look PFA, Lumsden MA, Templeton AA. Chromosome abnormalities in human embryos after in-vitro fertilization. Nature 1983;303: 336-8.

Magnuson T, Debrot S, Dimpfl J, Zweig A, Zamora T, Epstein CJ. The early lethality of autosomal monosomy in the mouse. J Exp Zoo1 1985;236:353-60. 46. Warburton D. The effect of maternal age on the frequency of trisomy: change in meiosis or in utero selection? In: Molecular and cytogenetic studies of nondisjunction. New York: Alan R Liss, 1989:165-81. 45.

Discussion DR. LUTHER M. T&BERT, Cary, North Carolina. This study is an important contribution to our understanding of the mechanisms and cause of fetal and embryonic loss. It is estimated that about 15% of all pregnancies end as clinical abortions. However, one carefully documented study showed that 61.9% of pregnancies are

lost before 12 weeks of gestation. Edmonds et al.’ studied 198 ovulatory cycles in a population of normal women who were attempting to become pregnant. One hundred eighteen women had positive test results for p-human chorionic gonadotropin between days 21 and 28 of their cycles (60% per ovulatory cycle). Sixty-seven conceptions (57%) were lost before the mother was aware of a pregnancy.’ A later study reported that 22% of 198 pregnancies in a group of healthy young women ended before pregnancy was clinically detected.* Although the percentages vary, it is well established that a large proportion of pregnancies are lost before recognition of a clinical pregnancy and that clinical abortions occur in an additional 15%. There can be no doubt that a large number of these pregnancy losses are the result of chromosomal errors. Estimates of the frequency of chromosomal abnormalities in spontaneously aborted tissues range from a high of 61% to a low of 30%, the variation probably because of differences in the sample being studied.s The frequency of chromosomal errors in preclinical abortions is unknown but may be inferred to be high. That most abnormal pregnancies are eliminated early is borne out by the finding that only 4.7% of pregnancies terminated by induced abortions beyond 7 weeks have chromosomal anomalies and that this number is further reduced to 0.53% at term, suggesting a remarkable propensity for the human female to eliminate abnormal conceptuses. Cytogenetic studies of human oocytes carried out in several different laboratories have confirmed that as many as 24% of oocytes obtained during IVF have chromosomal abnormalities, the large majority aneuploidies. Only 1% to 2% of oocytes have strnctural abnormalities. Aneuploidies in female gametes are not randomly distributed and have a higher than expected frequency in the D and G groups.’ Conversely, several studies have indicated that the incidence of structural abnormalities in sperm may be higher than that of aneuploidy and that chromosomal abnormalities in sperm are randomly distributed rather than being concentrated in specific chromosomal groupings6 The relatively high frequency of abnormalities in gametes of both sexes makes it likely that many of the chromosomal anomalies in embryos occur during meiosis, although some are clearly related to errors that occur in mitosis, as shown by the current study. By use of fluorescence in situ hybridization the, authors have devised a highly reproducible and accurate method for assessing numeric chromosomal abnormalities in developing embryos in vitro. Looking only at chromosomes X, Y, 18, and 16 they have shown an overall prevalence of 33% abnormal chromosome complements in the studied material. Morphologically normal embryos had a somewhat lower prevalence (23%), whereas embryos with morphologic anomalies had a prevalence of 42%. However, it must be remembered that only four pairs of chromosomes were studied and that the procedure used did not allow the detection of structural anomalies. Interestingly, mosaicism was the most commonly detected deviation from normal, con-

1200

Mum&

April 1995 Am J Obstet Gynecol

et al.

trasting with karyotypes on aborted material, where autosomal trisomy, sex chromosome monosomy, triploidy, and tetraploidy were the most common. If it is assumed that the prevalence of anomalies detected in this study reflects the situation in vivo, and it is probably a minimal estimate, then the reduction of chromosomal anomalies to 4.7% at 7 weeks and to 0.53% at term is truly a remarkable accomplishment. Bishop’ remarked in 1964 that “a considerable part of embryonic death is unavoidable and should be regarded as a normal way of eliminating unfit genotypes in each generation.” The rapidly growing ability to genetically assess the preimplantation embryo, not only chromosome numbers but also specific genes, with polymerase chain reaction or other methods will undoubtedly drastically change the practice of reproductive medicine in the coming years. I have two questions. (1) Does the high rate of mosaicism in the preimplantation embryos, as contrasted with a low frequency in aborted material, indicate that chromosomal mosaicism results in preclinical abortion, or could there be an alternative explanation? (2) Could you clarify the number of patients studied who were >40 years old? Were there enough in this group to skew the results toward a higher prevalence of. karyotypic abnormalities? REFERENCES 1. Edmonds

2. 3.

4.

5. 6.

7.

DK, Lindsay KS, Miller JK, Williamson E, Wood PJ. Early embryonic mortality in women. Fertil Steril 1982; 38:447-53. Wilcox AJ, Weinberg CR, O’Connor JF, et al. Incidence of early loss of pregnancy. New Engl J Med 1988;319:189-94. Hassold T, Chen N, Funkhouser J, et al. A cytogenetic study of 1000 spontaneous abortions. Ann Hum Genet 1980;44: 151-64. Burgoyne PS, Holland K, Stephens R. Incidence of numerical chromosome anomalies in human pregnancy: estimation from induced and spontaneous abortion data. Hum Reprod 1991;6:555-65. Pellestor F. Frequency and distribution of aneuploidy in human female gametes. Hum Genet 1991;86:283-8. Pellestor F. Differential distribution of aneuploidy in human gametes according to their sex. Hum Reprod 1991;6: 1252-8. Bishop MWH. Paternal contribution to embryonic death. J Reprod Fertil 1964;7:383-6.

DR. KAMRAN

MOGHISSI,

Detroit,

Michigan.

If your

pro-

cedure were to be used in conjunction with IVF, how much would it add to the cost of IVF? DR. HOWARD W. JONES, JR., Norfolk, Virginia. The authors have very nicely correlated age with regard to variations, but if we try to apply these data to situations such as preimplantation genetic diagnosis, we need to know certain other variables, for example, the meiotic status of the oocytes at aspiration, the number in the cohort, and the characteristics of the sperm used. If the authors could supply us with any information about these and other variables, we would feel much more comfortable in applying the informa-

tion to the younger people who are doing preimplantation studies. This work should be greatly facilitated by the current American Fertility Society ethics committee, which will approve experimentation on the preembryo. With the ethical cloud lifted, this work should proceed to give us some details about the other variables that have not been addressed so far. DR. JOHN E. BUSTER, Houston, Texas. I was particularly interested in your ideas about applying this to women > 40 years old underoing IVF to help select the optimal embryos for replacement. I wonder if you would comment on the possible added mortality that might occur from the invasion of the embryo by the biopsy and the increased culture time required to get to the requisite six or eight cells you need to take the blastomere out. Will you comment on the mortality rate that might be introduced by that added time versus the benefit of simply replacing embryos without the analysis? DR. SHERMAN ELIAS, Houston, Texas. I have a question regarding the spatial relationships of the chromosomespecific probes as they relate to aneuploidy detection. If there is only a single blastomere to analyze and two signals are seen, for example, for chromosome 18, how certain are you that a trisomy 18 is not being missed because of the spatial relationship where only signal could be superimposed on a second signal? In such a case, two signals would be seen under the microscope rather than three copies of chromosome 18. DR. ROSENWAKS (Closing). First, let me address the questions about the relative importance of mosaicism in this study. Mosaicism was the most common abnormality found. The importance of mosaicism in the early preimplantation embryo, particularly in relation to embryo survival, will likely depend on the number or proportion of chromosomally abnormal blastomeres. We believe that mosaic embryos containing a sizable number of normal cells may be able to develop normally, whereas those containing predominantly abnormal cells may be at a disadvantage. However, we have to emphasize that we are looking at very early embryos; few- data are available about mosaicism and its relative impact on embryo survival after IVF. Clearly, in this particular study, by analyzing not just one or two blastomeres but all the blastomeres, we were able to accurately assess the frequency of mosaicism, whereas other studies have not succeeded in obtaining this information. Thus mosaicism was found more often than we have previously appreciated. In this limited study only 64 embryos were studied from 42 women. There were very few patients >40 years old; only 16 embryos were studied from nine women >40 years old, and three embryos were abnormal. As to the question regarding the potential harm to embryos after 3 days of culture, we routinely culture embryos for 3 days, and there does not seem to be a problem with survival of embryos in our culture system.

Volume

172, Number Gynecol

4, Part 1

Mum6 et al.

1201

Am J Obstet

In fact, this is routine in our unit and the pregnancy rates and implantation rates per embryo are quite high in spite of the extra day of culture. As far as losing embryos with this procedure, only 1.5% of blastomeres were damaged during the analysis; there was another loss of about 2% of blastomeres at staining. There is a relatively high diagnostic efficiency with this procedure. Whereas there was a general overall fluorescence in situ hybridization failure rate of 12.2%, we lost very few embryos after embryo biopsies were performed. The comments of Dr. Jones in relation to the status of the gametes, both eggs and sperm, the relationship to ovarian stimulation per se, and the relation of the in vitro conditions, which may all have an impact on chromosomal makeup, must be. underscored because

we know that analysis of embryos from various clinics reveals different chromosomal abnormality rates. This may relate not only to stimulation protocols, gamete quality, and age of the woman but also to the conditions in the laboratory. The question of Dr. Elias about the accuracy of using one blastomere, and particularly of being sure that a diploid or trisomic cell is used, is a critical one. In other studies using three probes we found that the rate of signal overlap is only 0.8%. Clearly, from a single blastomere biopsy one cannot be as sure of the results as from assessment of two blastomeres. Up to this point we have not charged for the biopsy procedures. On the basis of the cost of the reagents and personnel, we anticipate an increased cost of $500 to $1000.