- Email: [email protected]
Clinical value of using an automated sperm morphology analyzer (IVOS)
MALE FACTOR
FERTILITY AND STERILITYt VOL. 71, NO. 2, FEBRUARY 1999 Copyright © 1999 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
Clinical value of using an automated sperm morphology analyzer (IVOS) Kevin Coetzee, M.Sc., Amanda de Villiers, Thinus F. Kruger, M.D., and Carl J. Lombard, Ph.D.* Reproductive Biology Unit, Department of Obstetrics and Gynaecology, Tygerberg Hospital, University of Stellenbosch, Tygerberg and Division of Epidemiology and Biostatistics, CERSA, Medical Research Council, Tygerberg, South Africa
Objective: To determine the clinical value of automated normal sperm morphology outcomes. Design: Prospective clinical study. Setting: Clinical and research assisted reproduction laboratory. Patient(s): Two hundred seven GIFT cycles. Intervention(s): The wife was induced to superovulate, laparoscopically aspirated, and the gametes were transferred laparoscopically. The husband’s sperm morphology was evaluated with use of a sperm morphology analyzer using the strict criteria classification system. Main Outcome Measure(s): Normal sperm morphology, IVF, and pregnancy outcomes. Result(s): The logistic regression model showed that normal sperm morphology was significantly associated with fertilization in vitro, as dependent (age) and independent variables. Analyzing the fertilization rates across the 5% normal sperm morphology cutoff point, a fertilization rate of 39.39% (#5%) compared with 62.92% (.5%) was obtained. The logistic regression model showed that normal sperm morphology was also a significant predictor of pregnancy when allowing for the number of oocytes transferred and female age. Analyzing the pregnancy rates across the 5% normal sperm morphology cutoff point, pregnancy rates of 15.15% (#5%) and 37.36% (.5%) were obtained. Conclusion(s): Normal sperm morphology as evaluated by the automated semen analyzer (IVOS) was shown to adhere to the same fertility cutoff point (5%), as determined by the manual evaluation of sperm morphology. Automated normal sperm morphology outcomes also were found to be significant predictors of IVF and pregnancy in a GIFT program. (Fertil Sterilt 1999;71:222–5. ©1999 by American Society for Reproductive Medicine.) Key Words: Normal sperm morphology, automated analysis, fertilization, pregnancies, gamete intrafallopian tube transfers
Received June 26, 1998; revised and accepted September 30, 1998. Reprint requests: Kevin Coetzee, M.Sc., Reproductive Biology Unit, 3rd floor, Tygerberg Hospital, Tygerberg 7505, South Africa (FAX: 027-21933 3084; E-mail: [email protected]). * IVOS, Hamilton Thorne Research, Beverly, Massachusetts. 0015-0282/99/$20.00 PII S0015-0282(98)00465-8
222
The strict criteria concept was conceived in the early 1980s on the basis of the homogeneous population of spermatozoa found at the level of the internal cervical os (1). This fortuitously coincided with the development of the in vitro laboratory at Tygerberg Hospital and, thereby, provided the situation in which to test whether the reasoning behind the principles of the strict criteria was sound. The first study to test the prognostic value of the normal sperm morphology outcomes, according to the strict criteria, obtained a fertilization rate of 82.5% and a pregnancy rate of 25.8% for the group with .14% normal forms (2). When the percentage of normal forms were #14%, the fertilization rate was 37.0%, and
no pregnancies were obtained. In a subsequent study a further normal sperm morphology fertility cutoff point (#4%, p-pattern) was identified (3). Since then, a number of international studies have confirmed the prognostic value of the 5% normal sperm morphology cutoff point (4). The universal adoption of standardized methodologies (i.e., slide and sample preparation and classification systems) is paramount to the achievement of comparable and consistent sperm morphology evaluations worldwide (5). The development of reliable automated sperm morphology evaluation systems, therefore, forms an important part of this strategy. The replacement of the visual evaluation of sperm
morphology with automated evaluations, however, first requires unequivocal proof of the system’s clinical predictive values. Unfortunately, no large prospective studies have been published in which automated normal sperm morphology outcomes have been used to measure the association of IVF and pregnancy with normal sperm morphology. In two previous studies Kruger et al. (6, 7) obtained promising results with the prediction of IVF. We, therefore, conducted this prospective study on 207 GIFT cycles to establish whether automated normal sperm morphology outcomes were predictive of IVF as well as pregnancy outcomes.
MATERIALS AND METHODS Materials An aliquot of the 207 semen samples to be used in GIFT procedures was obtained, and the slides for computerized evaluations were prepared (8). After complete liquefaction, the semen was washed once, using Ham’s F-10 medium (GIBCO BRL, Paisely, Scotland) supplemented with bovine serum albumin (0.003 g/mL; Seravac, Cape Town, South Africa), by centrifugation (350 3 g for 10 minutes). The resultant pellet, after the removal of the supernatant, was resuspended to obtain a concentration of 100 3 106 cells/ mL. A concentration-dependent droplet of semen sample was smeared thinly across cleaned slides and allowed to air dry (room temperature). The concentration, droplet size, and smear technique were standardized to produce 10 –20 sperm per high field magnification (5–10 sperm per computer screen). The densities required for computer evaluations are, therefore, approximately double that required for manual evaluations. The air-dried slides were stained with Diff-Quik (Merck Diagnostica, Darmstadt, Germany) stain according to a previously reported method (8). Briefly, the slides were fixed in solution 1 for 10 seconds, stained for 7 seconds in solution 2 and 7 seconds in solution 3, and washed in water. The Hamilton Thorne Research (IVOS, Hamilton Thorne Research, Beverly, MA) system is unique because it uses a “signature” method to evaluate the shape of the sperm head in addition to performing the standard axial measurements to determine normality. A valid cell’s signature (ellipse) extracted by a polar transformation method is compared with a normal signature using a least mean squared-based classifier (7). For a cell to be classified as normal, the parameters of size, shape, and acrosome must be within the normal limits. The computer settings were set as used at the individual laboratory for the evaluation of sperm morphology. Evaluations were performed with the use of 662-nm wavelength illumination in conjunction with a 3100 oil-immersion objective. Approximately 100 valid sperm cells were evaluated (blindly) per slide, and the percentage of normal sperm, as calculated by the computer, was recorded. FERTILITY & STERILITYt
Clomiphene citrate and hMG were used for ovarian hyperstimulation. Ovarian follicular growth was monitored with serial ultrasonographic measurements and serum LH determinations. hCG (10,000 IU) was administered when the dominant follicle reached a diameter of $18 mm, and follicular aspiration was performed 36 hours after administration. The oocytes were obtained by translaparoscopic follicle aspiration and graded according to their genetic (germinal vesicle, metaphase I, metaphase II) maturity. Semen samples were obtained by masturbation after 2– 4 days of sexual abstinence and allowed to liquefy. After liquefaction, the semen samples were prepared by means of the standard double wash (350 3 g for 10 minutes) and swim-up procedure. The sperm concentration used for the insemination (per oocyte) was determined by a normal sperm morphology evaluation performed before the GIFT procedure; 500,000 –1 3 106 cells (p-pattern, #4%) or 500,000 cells (g-pattern, 5%–14%) or 100,000 cells (normal pattern, .14%). Three to four metaphase II oocytes with the appropriate number of sperm cells were transferred laparoscopically through the fimbriated end of a (single) fallopian tube to a depth of 2 cm. All excess oocytes were inseminated, but only the outcomes of the metaphase II oocytes were considered in this study. Approximately 18 hours after insemination, the ova were observed for the presence of two distinct nucleoli containing pronuclei. Embryo cleavage was evaluated 40 hours after insemination. Pregnancy was confirmed by detecting increasing serum b-hCG concentration on days 12 and 16 after hCG administration. Clinical pregnancy was confirmed sonographically by the presence of a gestational sac at 7 weeks of pregnancy.
Statistics
The patient group in which the female age was #38 years was investigated for IVF and pregnancy outcomes. Normal sperm morphology was regarded and analyzed both as a continuous measurement and categorized at a cutoff point of 5%. The method of “Generalized Estimation Equations” was used as the regression model. The other limiting variables included in the logistic regression analysis were age (female) and number of oocytes transferred in the GIFT procedure (,4 or 4). Odds ratio and x2 analyses also were performed to determine differences across the 5% cutoff point.
RESULTS In Vitro Fertilization The logistic regression model showed that normal sperm morphology, as the only variable, was significantly (P 5 .0419) associated with IVF. When age was included in the model, the association was still found to be significant (P 5 .0317). Analyzing the fertilization rates across the 5% normal sperm morphology cutoff point, a fertilization rate of 39.39% (#5%) compared with 62.92% (.5%) was obtained 223
TABLE 1 In vitro fertilization and pregnancy rates according to automated semen morphology analysis outcomes. Cutoff points* #5%
.5%
2.7 (1.75)
17.1 (8.70)
0–5
6–44
39.39 (26/66)‡
62.92 (224/356)‡
15.15 (5/33)‡
37.36 (65/174)§
Variable Mean (6SD) morphology† Morphology range† Fertilization rate % (no. of oocytes) Pregnancy rate % (no. of cycles)
* Normal sperm morphology fertility cutoff points. † Percentage of normal sperm morphology. ‡ P 5 .0006. § P , .05.
(Table 1). The difference in fertilization rates was significant (P 5 .0006), even though no adjustment for age was made.
Pregnancy The logistic regression model showed that normal sperm morphology was a significant (odds ratio 5 3.39; P 5 .0210) predictor of pregnancy when allowing for the number of oocytes transferred and age. Analyzing the pregnancy rates across the 5% normal sperm morphology cutoff point, pregnancy rates of 15.15% (#5%) and 37.36% (.5%) were obtained (Table 1). The difference in pregnancy rates was significant (odds ratio 3.34; P , .05), even though no adjustment for number of oocytes transferred and age were made. The pregnancy rates above the 5% cutoff point remained relatively constant.
DISCUSSION Within its limitations, normal sperm morphology now has been recognized as a simple yet valuable prognostic and diagnostic variable in assisted reproduction programs. However, to entrench the use of normal sperm morphology outcomes worldwide, a concerted effort will have to be made to standardize the evaluation methodologies used. The development of automated semen analyzers may facilitate this standardization strategy, thereby making the value of normal sperm morphology evaluations available to more assisted reproduction centers worldwide. The Hamilton Thorne Research IVOS is one such automated system currently being developed. In the first clinical trial conducted, using the IVOS, the examination of the association between automated normal sperm morphology outcomes and IVF yielded promising results (9). Normal sperm morphology was significantly associated with IVF. The fertilization rates categorized according to normal sperm 224
Coetzee et al.
Automated sperm morphology analysis
morphology fertility groups were 46% in the p-pattern group, 73% in the 5%–9% group, 82% in the 10%–14% group, and 85% in the .14% group. The judgment of these systems cannot rest on a single trial; it requires the performance of a number of independent prospective clinical studies to confirm the predictive value of automated normal sperm morphology outcomes. Therefore, a further study was performed in which 207 GIFT cycles were evaluated prospectively for fertilization (P 5 .0006) and pregnancy (P , .05) outcomes. The logistic regression model applied in the study found normal sperm morphology, as evaluated by the automated system, to be associated significantly with both IVF and pregnancy. This outcome was true for both univariate and multivariate models. Categorizing normal sperm morphology outcomes also reaffirmed the significance of the 5% normal sperm morphology fertility cutoff point, as determined by the manual evaluation of sperm morphology. The fertilization rates across this cutoff point were 39.4% (#5%) and 62.9% (.5%), whereas the pregnancy rates across the cutoff point were 15.2% (#5%) compared with 37.36% (.5%). The association between fertilization and normal sperm morphology observed in this latest study is markedly similar to that obtained in the study performed by Kruger et al. (9). It was speculated in that study that because of the differences between manual and automated evaluations (intervariation) and finding that all nonfertilizations occurred in the ,10% group that new normal sperm morphology fertility thresholds may be intrinsic for automated evaluations. The evaluation of our results showed no evidence of the presentation of new thresholds. In another study, using the IVOS, it also was shown that the particular automated system’s repeatability (accuracy) compares with that of an experienced observer evaluating sperm morphology manually (10). In light of the accuracy, repeatability, and predictive value of the automated normal sperm morphology outcomes obtained with the IVOS, they may be used with confidence in deciding the treatment regiment of infertile couples by identifying patients with lower IVF potential. The accuracy when using the system is, however, subject to the quality of the slides produced (sample and slide preparation) and the technologist’s understanding of and experience in using the automated system. By adhering to prescribed quality control standards, the Hamilton Thorne Research system can be applied for the routine evaluation of sperm morphology. References 1. Menkveld R, Stander FSH, Kotze TJvW, Kruger TF, van Zyl JA. The evaluation of morphological characteristics of human spermatozoa according to stricter criteria. Hum Reprod 1990;5:586 –92. 2. Kruger TF, Menkveld R, Stander FSH, Lombard CJ, van der Merwe JP, van Zyl JA, et al. Sperm morphological features as a prognostic factor in in vitro fertilization. Fertil Steril 1986;46:1118 –23. 3. Kruger TF, Acosta AA, Simmons KF, Swanson RJ, Matta JF, Oehninger S. Predictive value of abnormal sperm morphology in in vitro fertilization. Fertil Steril 1988;49:112–7.
Vol. 71, No. 2, February 1999
4. Coetzee K, Kruger TF, Lombard CJ. Predictive value of normal sperm morphology: a structured review. Hum Reprod Update 1998;4:73– 82. 5. Ombelet W, Pollet H, Bosmans E, Vereecken A. Results of a questionnaire on sperm morphology assessment. Hum Reprod 1997;12: 1015–20. 6. Kruger TF, Du Toit TC, Franken DR, Acosta AA, Oehninger SC, Menkveld R, et al. A new computerized method of reading sperm morphology (strict criteria) is as efficient as technician reading. Fertil Steril 1993;59:202–9. 7. Kruger TF, du Toit TC, Franken DR, Menkveld R, Lombard CJ. Sperm morphology: assessing the agreement between the manual method (strict criteria) and the sperm morphology analyzer IVOS. Fertil Steril 1995;63:134 – 41.
FERTILITY & STERILITYt
8. Lacquet FA, Kruger TF, du Toit TC, Lombard CJ, Sanchez Sarmiento CA, de Villiers A, et al. Slide preparation and staining procedures for reliable results using computerized morphology (IVOS). Arch Androl 1996;36:133– 8. 9. Kruger TF, Lacquet FA, Sanchez Sarmiento CA, Menkveld R, Ozgur K, Lombard CJ, et al. A prospective study on the predictive value of normal sperm morphology as evaluated by computer (IVOS). Fertil Steril 1996;66:285–91. 10. Coetzee K, Kruger TF, Lombard CJ. Repeatability and variance analysis on multiple computer-assisted (IVOS) sperm morphology readings. Andrologia (In press).
225
FERTILITY AND STERILITYt VOL. 71, NO. 2, FEBRUARY 1999 Copyright © 1999 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
Clinical value of using an automated sperm morphology analyzer (IVOS) Kevin Coetzee, M.Sc., Amanda de Villiers, Thinus F. Kruger, M.D., and Carl J. Lombard, Ph.D.* Reproductive Biology Unit, Department of Obstetrics and Gynaecology, Tygerberg Hospital, University of Stellenbosch, Tygerberg and Division of Epidemiology and Biostatistics, CERSA, Medical Research Council, Tygerberg, South Africa
Objective: To determine the clinical value of automated normal sperm morphology outcomes. Design: Prospective clinical study. Setting: Clinical and research assisted reproduction laboratory. Patient(s): Two hundred seven GIFT cycles. Intervention(s): The wife was induced to superovulate, laparoscopically aspirated, and the gametes were transferred laparoscopically. The husband’s sperm morphology was evaluated with use of a sperm morphology analyzer using the strict criteria classification system. Main Outcome Measure(s): Normal sperm morphology, IVF, and pregnancy outcomes. Result(s): The logistic regression model showed that normal sperm morphology was significantly associated with fertilization in vitro, as dependent (age) and independent variables. Analyzing the fertilization rates across the 5% normal sperm morphology cutoff point, a fertilization rate of 39.39% (#5%) compared with 62.92% (.5%) was obtained. The logistic regression model showed that normal sperm morphology was also a significant predictor of pregnancy when allowing for the number of oocytes transferred and female age. Analyzing the pregnancy rates across the 5% normal sperm morphology cutoff point, pregnancy rates of 15.15% (#5%) and 37.36% (.5%) were obtained. Conclusion(s): Normal sperm morphology as evaluated by the automated semen analyzer (IVOS) was shown to adhere to the same fertility cutoff point (5%), as determined by the manual evaluation of sperm morphology. Automated normal sperm morphology outcomes also were found to be significant predictors of IVF and pregnancy in a GIFT program. (Fertil Sterilt 1999;71:222–5. ©1999 by American Society for Reproductive Medicine.) Key Words: Normal sperm morphology, automated analysis, fertilization, pregnancies, gamete intrafallopian tube transfers
Received June 26, 1998; revised and accepted September 30, 1998. Reprint requests: Kevin Coetzee, M.Sc., Reproductive Biology Unit, 3rd floor, Tygerberg Hospital, Tygerberg 7505, South Africa (FAX: 027-21933 3084; E-mail: [email protected]). * IVOS, Hamilton Thorne Research, Beverly, Massachusetts. 0015-0282/99/$20.00 PII S0015-0282(98)00465-8
222
The strict criteria concept was conceived in the early 1980s on the basis of the homogeneous population of spermatozoa found at the level of the internal cervical os (1). This fortuitously coincided with the development of the in vitro laboratory at Tygerberg Hospital and, thereby, provided the situation in which to test whether the reasoning behind the principles of the strict criteria was sound. The first study to test the prognostic value of the normal sperm morphology outcomes, according to the strict criteria, obtained a fertilization rate of 82.5% and a pregnancy rate of 25.8% for the group with .14% normal forms (2). When the percentage of normal forms were #14%, the fertilization rate was 37.0%, and
no pregnancies were obtained. In a subsequent study a further normal sperm morphology fertility cutoff point (#4%, p-pattern) was identified (3). Since then, a number of international studies have confirmed the prognostic value of the 5% normal sperm morphology cutoff point (4). The universal adoption of standardized methodologies (i.e., slide and sample preparation and classification systems) is paramount to the achievement of comparable and consistent sperm morphology evaluations worldwide (5). The development of reliable automated sperm morphology evaluation systems, therefore, forms an important part of this strategy. The replacement of the visual evaluation of sperm
morphology with automated evaluations, however, first requires unequivocal proof of the system’s clinical predictive values. Unfortunately, no large prospective studies have been published in which automated normal sperm morphology outcomes have been used to measure the association of IVF and pregnancy with normal sperm morphology. In two previous studies Kruger et al. (6, 7) obtained promising results with the prediction of IVF. We, therefore, conducted this prospective study on 207 GIFT cycles to establish whether automated normal sperm morphology outcomes were predictive of IVF as well as pregnancy outcomes.
MATERIALS AND METHODS Materials An aliquot of the 207 semen samples to be used in GIFT procedures was obtained, and the slides for computerized evaluations were prepared (8). After complete liquefaction, the semen was washed once, using Ham’s F-10 medium (GIBCO BRL, Paisely, Scotland) supplemented with bovine serum albumin (0.003 g/mL; Seravac, Cape Town, South Africa), by centrifugation (350 3 g for 10 minutes). The resultant pellet, after the removal of the supernatant, was resuspended to obtain a concentration of 100 3 106 cells/ mL. A concentration-dependent droplet of semen sample was smeared thinly across cleaned slides and allowed to air dry (room temperature). The concentration, droplet size, and smear technique were standardized to produce 10 –20 sperm per high field magnification (5–10 sperm per computer screen). The densities required for computer evaluations are, therefore, approximately double that required for manual evaluations. The air-dried slides were stained with Diff-Quik (Merck Diagnostica, Darmstadt, Germany) stain according to a previously reported method (8). Briefly, the slides were fixed in solution 1 for 10 seconds, stained for 7 seconds in solution 2 and 7 seconds in solution 3, and washed in water. The Hamilton Thorne Research (IVOS, Hamilton Thorne Research, Beverly, MA) system is unique because it uses a “signature” method to evaluate the shape of the sperm head in addition to performing the standard axial measurements to determine normality. A valid cell’s signature (ellipse) extracted by a polar transformation method is compared with a normal signature using a least mean squared-based classifier (7). For a cell to be classified as normal, the parameters of size, shape, and acrosome must be within the normal limits. The computer settings were set as used at the individual laboratory for the evaluation of sperm morphology. Evaluations were performed with the use of 662-nm wavelength illumination in conjunction with a 3100 oil-immersion objective. Approximately 100 valid sperm cells were evaluated (blindly) per slide, and the percentage of normal sperm, as calculated by the computer, was recorded. FERTILITY & STERILITYt
Clomiphene citrate and hMG were used for ovarian hyperstimulation. Ovarian follicular growth was monitored with serial ultrasonographic measurements and serum LH determinations. hCG (10,000 IU) was administered when the dominant follicle reached a diameter of $18 mm, and follicular aspiration was performed 36 hours after administration. The oocytes were obtained by translaparoscopic follicle aspiration and graded according to their genetic (germinal vesicle, metaphase I, metaphase II) maturity. Semen samples were obtained by masturbation after 2– 4 days of sexual abstinence and allowed to liquefy. After liquefaction, the semen samples were prepared by means of the standard double wash (350 3 g for 10 minutes) and swim-up procedure. The sperm concentration used for the insemination (per oocyte) was determined by a normal sperm morphology evaluation performed before the GIFT procedure; 500,000 –1 3 106 cells (p-pattern, #4%) or 500,000 cells (g-pattern, 5%–14%) or 100,000 cells (normal pattern, .14%). Three to four metaphase II oocytes with the appropriate number of sperm cells were transferred laparoscopically through the fimbriated end of a (single) fallopian tube to a depth of 2 cm. All excess oocytes were inseminated, but only the outcomes of the metaphase II oocytes were considered in this study. Approximately 18 hours after insemination, the ova were observed for the presence of two distinct nucleoli containing pronuclei. Embryo cleavage was evaluated 40 hours after insemination. Pregnancy was confirmed by detecting increasing serum b-hCG concentration on days 12 and 16 after hCG administration. Clinical pregnancy was confirmed sonographically by the presence of a gestational sac at 7 weeks of pregnancy.
Statistics
The patient group in which the female age was #38 years was investigated for IVF and pregnancy outcomes. Normal sperm morphology was regarded and analyzed both as a continuous measurement and categorized at a cutoff point of 5%. The method of “Generalized Estimation Equations” was used as the regression model. The other limiting variables included in the logistic regression analysis were age (female) and number of oocytes transferred in the GIFT procedure (,4 or 4). Odds ratio and x2 analyses also were performed to determine differences across the 5% cutoff point.
RESULTS In Vitro Fertilization The logistic regression model showed that normal sperm morphology, as the only variable, was significantly (P 5 .0419) associated with IVF. When age was included in the model, the association was still found to be significant (P 5 .0317). Analyzing the fertilization rates across the 5% normal sperm morphology cutoff point, a fertilization rate of 39.39% (#5%) compared with 62.92% (.5%) was obtained 223
TABLE 1 In vitro fertilization and pregnancy rates according to automated semen morphology analysis outcomes. Cutoff points* #5%
.5%
2.7 (1.75)
17.1 (8.70)
0–5
6–44
39.39 (26/66)‡
62.92 (224/356)‡
15.15 (5/33)‡
37.36 (65/174)§
Variable Mean (6SD) morphology† Morphology range† Fertilization rate % (no. of oocytes) Pregnancy rate % (no. of cycles)
* Normal sperm morphology fertility cutoff points. † Percentage of normal sperm morphology. ‡ P 5 .0006. § P , .05.
(Table 1). The difference in fertilization rates was significant (P 5 .0006), even though no adjustment for age was made.
Pregnancy The logistic regression model showed that normal sperm morphology was a significant (odds ratio 5 3.39; P 5 .0210) predictor of pregnancy when allowing for the number of oocytes transferred and age. Analyzing the pregnancy rates across the 5% normal sperm morphology cutoff point, pregnancy rates of 15.15% (#5%) and 37.36% (.5%) were obtained (Table 1). The difference in pregnancy rates was significant (odds ratio 3.34; P , .05), even though no adjustment for number of oocytes transferred and age were made. The pregnancy rates above the 5% cutoff point remained relatively constant.
DISCUSSION Within its limitations, normal sperm morphology now has been recognized as a simple yet valuable prognostic and diagnostic variable in assisted reproduction programs. However, to entrench the use of normal sperm morphology outcomes worldwide, a concerted effort will have to be made to standardize the evaluation methodologies used. The development of automated semen analyzers may facilitate this standardization strategy, thereby making the value of normal sperm morphology evaluations available to more assisted reproduction centers worldwide. The Hamilton Thorne Research IVOS is one such automated system currently being developed. In the first clinical trial conducted, using the IVOS, the examination of the association between automated normal sperm morphology outcomes and IVF yielded promising results (9). Normal sperm morphology was significantly associated with IVF. The fertilization rates categorized according to normal sperm 224
Coetzee et al.
Automated sperm morphology analysis
morphology fertility groups were 46% in the p-pattern group, 73% in the 5%–9% group, 82% in the 10%–14% group, and 85% in the .14% group. The judgment of these systems cannot rest on a single trial; it requires the performance of a number of independent prospective clinical studies to confirm the predictive value of automated normal sperm morphology outcomes. Therefore, a further study was performed in which 207 GIFT cycles were evaluated prospectively for fertilization (P 5 .0006) and pregnancy (P , .05) outcomes. The logistic regression model applied in the study found normal sperm morphology, as evaluated by the automated system, to be associated significantly with both IVF and pregnancy. This outcome was true for both univariate and multivariate models. Categorizing normal sperm morphology outcomes also reaffirmed the significance of the 5% normal sperm morphology fertility cutoff point, as determined by the manual evaluation of sperm morphology. The fertilization rates across this cutoff point were 39.4% (#5%) and 62.9% (.5%), whereas the pregnancy rates across the cutoff point were 15.2% (#5%) compared with 37.36% (.5%). The association between fertilization and normal sperm morphology observed in this latest study is markedly similar to that obtained in the study performed by Kruger et al. (9). It was speculated in that study that because of the differences between manual and automated evaluations (intervariation) and finding that all nonfertilizations occurred in the ,10% group that new normal sperm morphology fertility thresholds may be intrinsic for automated evaluations. The evaluation of our results showed no evidence of the presentation of new thresholds. In another study, using the IVOS, it also was shown that the particular automated system’s repeatability (accuracy) compares with that of an experienced observer evaluating sperm morphology manually (10). In light of the accuracy, repeatability, and predictive value of the automated normal sperm morphology outcomes obtained with the IVOS, they may be used with confidence in deciding the treatment regiment of infertile couples by identifying patients with lower IVF potential. The accuracy when using the system is, however, subject to the quality of the slides produced (sample and slide preparation) and the technologist’s understanding of and experience in using the automated system. By adhering to prescribed quality control standards, the Hamilton Thorne Research system can be applied for the routine evaluation of sperm morphology. References 1. Menkveld R, Stander FSH, Kotze TJvW, Kruger TF, van Zyl JA. The evaluation of morphological characteristics of human spermatozoa according to stricter criteria. Hum Reprod 1990;5:586 –92. 2. Kruger TF, Menkveld R, Stander FSH, Lombard CJ, van der Merwe JP, van Zyl JA, et al. Sperm morphological features as a prognostic factor in in vitro fertilization. Fertil Steril 1986;46:1118 –23. 3. Kruger TF, Acosta AA, Simmons KF, Swanson RJ, Matta JF, Oehninger S. Predictive value of abnormal sperm morphology in in vitro fertilization. Fertil Steril 1988;49:112–7.
Vol. 71, No. 2, February 1999
4. Coetzee K, Kruger TF, Lombard CJ. Predictive value of normal sperm morphology: a structured review. Hum Reprod Update 1998;4:73– 82. 5. Ombelet W, Pollet H, Bosmans E, Vereecken A. Results of a questionnaire on sperm morphology assessment. Hum Reprod 1997;12: 1015–20. 6. Kruger TF, Du Toit TC, Franken DR, Acosta AA, Oehninger SC, Menkveld R, et al. A new computerized method of reading sperm morphology (strict criteria) is as efficient as technician reading. Fertil Steril 1993;59:202–9. 7. Kruger TF, du Toit TC, Franken DR, Menkveld R, Lombard CJ. Sperm morphology: assessing the agreement between the manual method (strict criteria) and the sperm morphology analyzer IVOS. Fertil Steril 1995;63:134 – 41.
FERTILITY & STERILITYt
8. Lacquet FA, Kruger TF, du Toit TC, Lombard CJ, Sanchez Sarmiento CA, de Villiers A, et al. Slide preparation and staining procedures for reliable results using computerized morphology (IVOS). Arch Androl 1996;36:133– 8. 9. Kruger TF, Lacquet FA, Sanchez Sarmiento CA, Menkveld R, Ozgur K, Lombard CJ, et al. A prospective study on the predictive value of normal sperm morphology as evaluated by computer (IVOS). Fertil Steril 1996;66:285–91. 10. Coetzee K, Kruger TF, Lombard CJ. Repeatability and variance analysis on multiple computer-assisted (IVOS) sperm morphology readings. Andrologia (In press).
225