1 Department of Growth and Reproduction, Copenhagen University Hospital, DK-2100 Copenhagen, 2 Department of Biostatistics, University of Copenhagen, Panum Institute, DK-2200 Copenhagen, 3 Department of Occupational Medicine, Aarhus University Hospital, DK-8000 Aarhus, Denmark and 4 Department of Urology, Malmö University Hospital, Malmö, Sweden
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: computer-assisted semen analysis/male fertility/semen quality/sperm concentration
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In an attempt to make the assessments of semen quality more objective and detailed, tools for computer-assisted semen analysis (CASA) have been developed. By use of CASA several specific motility parameters describing the movements of spermatozoa in a more detailed manner can be obtained. Furthermore, the classification into motile and immotile spermatozoa can be based on well-defined velocity thresholds.
Parallel with attempts to improve the technical performance of CASA systems, it is important to investigate the biological relevance of CASA parameters in the context of prediction of male fertility potential, a knowledge which is of crucial importance for understanding the biology of fertilization and also for diagnosis and treatment of male infertility.
Several studies have addressed the issue of CASA parameters as predictors of male fertility (Aitken et al., 1984; Barratt et al., 1993
; Irvine et al., 1994
; Paston et al., 1994
; Krause, 1995
). However, those studies have often been based on couples recruited from infertility clinics. In such studies the most obvious cases of female infertility, including tubular occlusion or anovulation, can be exlcluded. On the other hand, the percentage of couples with unexplained infertility, which may also be caused by a female factor, is quite significant and men with normal semen parameters may be rather underrepresented. Furthermore, frequently the outcome of different types of artificial reproduction procedures was used as an end-point, making the results of more limited relevance for the naturally achieved pregnancies. The use of such different approaches and lack of standardization in the field of CASA may explain the fact that virtually all CASA parameters have been reported to be correlated with male fertility potential in at least one study set-up. Taking men from the general population is advantageous because they will span the spectrum of semen profiles from very poor to very good. However, since the female partner is not investigated thoroughly, some men will be misclassified as infertile even though the infertility of the couple is almost entirely down to the female partner. This phenomenon may introduce some `noise' in the final statistical analysis. Information on the relationship between CASA parameters and the chance of conception in couples from a general population is very sparse.
Recently, it was shown that conventionally assessed sperm concentration is a strong predictor of fertility in normal males (Bonde et al., 1998a). The objective of the present study was to investigate whether CASA motility parameters may add to the information regarding semen quality in relation to fertility of men recruited from the general population.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Computer-aided semen analysis (CASA)
The CASA system employed in this study was the CRISMAS system, version 1.0 (Image House A/S, Copenhagen, Denmark). It calculated CASA parameters of sperm motility according to the guidelines from WHO (1992). In addition to these specific motility parameters, the computer also calculated sperm concentration, total sperm count, the percentage of motile spermatozoa in the sample, and a classification of the spermatozoa into immotile (curvilinear velocity, VCL <5 µm/s), locally motile (VCL = 525 µm/s) and motile (VCL >25 µm/s) was made. CRISMAS had an edit function that allowed the user to correct wrong sperm tracks calculated by the computer. It was thus possible to correct misinterpretations such as other particles being recognized as immotile spermatozoa or wrong tracking resulting from paths crossing each other. This property minimized the error rate and made the results more valid. The main components of the CRISMAS system were a Nikon microscope with positive phase contrast optics, a charge-coupled device (CCD) video camera and a personal computer with a digital frame grabber and CRISMAS image processing software. The measuring time for CRISMAS was 4x2.6 s and the frame sampling frequency was 25 Hz.
Experimental conditions
CASA was made on semen samples collected at start of the enrolment, and the analysis is based on samples from 358 men of the 419 men included in the whole study. Thirty samples collected in a pilot study in 1992 were not stored on videotapes for analysis by CRISMAS. Eight samples showed azoospermia (n = 7) or severe oligozoospermia (n = 1); one sample was too viscous to analyse; 12 samples were not analysed for different but absolutely random causes; 10 samples could not be analysed by CRISMAS because of a low quality of the video recordings of these samples resulting from an insufficient dilution. This could mean a problem of selection bias if the sperm concentration of these samples was higher than the other samples. Attempts were made to solve this problem by including the manually determined sperm concentrations from these 10 samples in the statistical analyses. This did not change the results.
The semen samples were collected and recorded on videotapes for CASA analysis in both Copenhagen and Aarhus, and the same recording protocols were applied in the two centres except of the sample volume pipetted into Makler chamber being 2.5 µl in Copenhagen and 4.5 µl in Århus.
Freshly collected semen samples were allowed to liquefy for 20 min at 37°C. They were diluted appropriately in mixed agglutination reaction (MAR) test buffer (9 mmol/l KH2PO4, 28 mmol/l Na2HPO4, 11 mmol/l NaCl) and the diluted sample was pipetted into a Makler chamber, which was placed on a heated microscope stage (37°C). Video recordings were made from four different fields of the chamber using a x20 magnification objective on the microscope. CASA analysis was based upon capturing sequences of 64 frames per field and counting a minimum of 100 spermatozoa.
Validation of sperm concentration assessment by CRISMAS
In order to evaluate the accuracy of CRISMAS in distinguishing non-motile spematozoa from other particles, for the 358 samples included in the study, sperm concentration values obtained by CASA were compared to the results obtained by counting in a BürkerTürk chamber, according to the WHO guidelines (1992). The quality of the standard sperm concentration assessment was regularly checked by internal quality assurance programme showing an inter-technician coefficient of variation of <15%.
Evaluation of the effect of dilution
For ten randomly selected ejaculates, CASA analysis was performed on undiluted samples as well as samples diluted according to the above-mentioned protocol. The results of the two assessments were compared to each other.
Statistical methods
CRISMAS calculates CASA parameters for all spermatozoa and for only the subpopulation of motile spermatozoa in a semen sample, respectively. The following statistical analyses were made with CASA parameters for the motile spermatozoa only, because it was found that this sperm population is most relevant in the context of fertilization. When motility parameters for all spermatozoa were utilized, the results of the statistical analyses were virtually the same (data not shown).
The data were analysed in univariate and multivariate discrete Cox regression analyses. The regression model was validated by standard regression techniques. By this regression method the potential of the CASA parameters to predict the chance of pregnancy was examined. All the variables were entered as linear covariates, except for sperm concentration, total number of spermatozoa, concentration of motile spermatozoa and straight line velocity (VSL), which were entered as logarithmic values in order to fulfil the assumption of linearity of the covariates. Model selection was done by backwards model selection techniques and combined with an initial model selection strategy based on subject matter knowledge.
In the multivariate analyses the results were adjusted for a possible confounding effect of several male and female factors. These factors were: abstinence time (number of days since last ejaculation), occupation (1 = metal worker, 2 = office worker, 3 = nurse, 4 = daycare worker), oral contraception as last contraceptive method (1 = yes, 0 = no), female age (years), female smoking (1 = smoking, 0 = not smoking), female body mass index (BMI; kg/m2), if the woman had ever experienced diseases in the internal female reproductive organs (1 = yes, 0 = no) and length of menstrual cycle (days). Because the video recording procedure differed between the two centres the results were also controlled for a confounding effect of `centre' (1 = Aarhus, 2 = Copenhagen).
Separate Cox regression analyses were done with a subpopulation of spermatozoa having motility characteristics reported to be required for penetration of cervical mucus (`mucus-penetrating spermatozoa'). These analyses were done in order to clarify whether the results were skewed by inclusion of spermatozoa not capable of penetrating cervical mucus and which can be considered as not physiologically relevant for the process of fertilization. Mucus-penetrating spermatozoa have been reported to be characterized by minimum values of VSL, VSL/VAP = straightness and amplitude of lateral head displacement (ALH) of 25 µm/s, 90% and 2.5 µm/s, respectively (Mortimer, 1994).
MannWhitney test was applied in order to evaluate the effect of dilution on CASA parameters. The statistical analyses were carried out using the Statistical Analysis System (SAS) for personal computers, version 6.12 for Windows (SAS Institute Inc., Cary, NC, USA).
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
CASA parameters and fertility
The results of univariate Cox regression analyses are shown in Table III. Ln(sperm concentration), ln (total sperm count), ln(concentration of motile spermatozoa) and percentage of motile spermatozoa were most significantly associated with the probability of achieving a pregnancy, but VCL, average path velocity (VAP), STR and BCF were also of significant predictive value for fertility.
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The finding of concentration of motile spermatozoa as independent and the most significant parameter predictive for fertility outcome, led to the question of whether CASA is of any aid in assessment of male reproductive function. However, it should be kept in mind that use of CASA assessment is more objective and reproducible than technician-based motility assessment. In a previous analysis, based on the same material (Bonde et al., 1998a), conventionally performed motility assessment was not an independent predictor of fertility. Recently published `Guidelines on the application of CASA technology in the analysis of spermatozoa' (ESHRE Andrology Special Interest Group, 1998
) do not recommend use of such equipment for determination of the proportion of motile spermatozoa, due to lack of reliability in distinguishing between debris and dead spermatozoa. However, the current study showed that by using CRISMAS, median overestimation of concentration was 1.6x106/ml, which is <5%.
Furthermore, it cannot be excluded that more refined analyses, based on well-defined subpopulations of spermatozoa, may disclose other motility parameters as predictors of male fertility. In this study a separate analysis was performed on a subtype of spermatozoa possessing characteristics previously reported as necessary for cervical mucus penetration (VSL >25 µm/s, STR >90% and ALH 2.5 µm) (Mortimer, 1994). This analysis did not change the conclusion of the study. However, the studies made by Mortimers group were based on use of HamiltonThorne Motility AnalyzerIntegrated Visual Optics System (HTMA-IVOS) equipment and due to differences in settings and technical performance the subpopulation characteristics may not be transferable from one instrument to another.
Previous studies have pointed out other motility parameters as being of significance for predicting male fertility. Thus, Barrat et al. (1993) found the total number of spermatozoa and VAP to be predictors of chance of pregnancy, whereas Irvine et al. (1994) found mean head area and the manually determined percentage of progressively motile spermatozoa as being significant covariates (Irvine et al., 1994). Other parameters as ALH, VSL, VCL and linearity (LIN) have also been reported to be correlated with fertility (Aitken et al., 1984
; Barratt et al., 1993
; Irvine et al., 1994
; Paston et al., 1994
; Krause, 1995
). However, all these studies were based on patients from infertility clinics, which imply the above-mentioned limitations in extrapolating the data to the general population. The current study also found several of those motility parameters to be significantly correlated with fertility outcome when a univariate Cox regression model was applied. However, when entered in a multivariate regression model only sperm concentration and concentration as well as percentage motile spermatozoa were found to be independent fertility predictors. On the other hand,one cannot exclude that for some of the parameters the lack of significance in the multivariate model is due to a type II error. Thus, the level of statistical significance might be reached if the number of individuals included were increased. Furthermore, it cannot be excluded that a different model may have been produced if a forwards selection strategy was used instead of the backwards selection method used in this study.
It should also be taken into consideration that the most recent recommendations for use of CASA (ESHRE Andrology Special Interest Group, 1998) state that only kinematic data acquired at
50 Hz is reliable. However, according to a recently published study by Mortimer and Swan (1999), the change from 25 to 50 Hz, for non-hyperactivated spermatozoa, has no effect on VSL, VAP, ALH and implies only a slight increase in VCL. It seems, therefore, unlikely that the major conclusions of this study would change if a higher image sampling frequency were used.
In this study a recently developed system (CRISMAS) was used for CASA analysis. It could be questioned whether the results of this study have any relevance for users of other computer systems. However, in a study of 234 infertility couples Krause (1995) also found sperm concentration and percentage motile spermatozoa to be predictors of fertility outcome in vivo (Krause, 1995). Interestingly, using the HamiltonThorne CASA system in this study the regression coefficients found for the covariates of ln(sperm concentration) and percentage motile spermatozoa were very similar to those found in the current study (0.268 versus 0.234 and 0.010 versus 0.012, respectively).
VCL >25 µm/s was used as the threshold value for the category of motile spermatozoa, whereas in some other papers this definition is based on the VSL or VCL value. Neither the WHO manual (1992) nor the guidelines from the ESHRE Andrology Special Interest Group (1996, 1998) give any scientific the or recommendations on which of these criteria is most significant from a biological point of view. As already pointed out by others (Lenzi, 1997; ESHRE Andrology Special Interest Group, 1998
), there is an urgent need for standardization of CASA systems, particularly as many of the motility parameters are calculated from a smoothed average path and there is no agreement on a common smoothing algorithm.
Samples with a high concentration of spermatozoa were diluted with a MAR test buffer in order to facilitate the CASA analysis. In order to minimize the impact of the dilution procedure, the use of autologous sperm-free seminal plasma as dilutant has been recommended. A comparison of diluted and undiluted samples showed no impact of MAR test buffer on the percentage of motile spermatozoa or motility parameters, except BCF and STR, in the population of motile spermatozoa.
Increasing concentration of spermatozoa as assessed by CRISMAS was positively correlated with fertility outcome up to the level of ~30 x106/ml for motile spermatozoa and slightly higher for the total population of spermatozoa. This finding is in agreement with recently published results of manually determined sperm concentration assessment based on the same population material (Bonde et al., 1998a). An identical threshold value for the manual and CRISMAS assessment might indicate that sperm concentration values obtained by the CASA system used in this study are in good agreement with the traditional sperm counting methods, which are also in the low concentration range. From the biological point of view it seems interesting that increasing sperm concentration, up to a certain value, implies a higher fertility whereas for percentage motile cells such a threshold value does not exist.
One of the limitations of this study is the fact that the analysis is based on only one sample from each of the 358 men. Due to intra-individual variation (WHO, 1992), it is possible that the semen sample delivered when the couple entered the study may not express semen characteristics at the time when the conception took place, which may have occurred up to 6 months after the ejaculate was collected. A recently published study based on the same material addressed this problem in relation to sperm concentration. This analysis has shown that using sperm concentration in a semen sample delivered at the time of beginning of the relevant menstrual cycle did not imply better prediction of cycle-specific fertilization rate, as compared to the ejaculate produced at the beginning of the study (Bonde et al., 1998a). The intra-individual variation in motility parameters is not higher than that found for sperm concentration. Thus, although it is not possible to prove this, the results of the current study might not change significantly even if `cycle specific' motility parameters were available.
In conclusion, in this study of 358 first pregnancy planners it was found that the concentration of spermatozoa with VCL >25 µm/s was the most significant and independent CASA parameter in predicting male fertility potential. This appears to be the first study showing a correlation of CASA parameters of men from the general population with their fertility in vivo.
![]() |
Acknowledgments |
---|
![]() |
Notes |
---|
* The Danish First Pregnancy Planner Study is a collaborative follow-up study on environmental and biological determinants of fertility. The project is co-ordinated by the Steno Institute of Public Health, University of Århus and is undertaken in Collaboration with the Department of Growth and Reproduction, National University Hospital in Copenhagen. The team includes Jens Peter E.Bonde, Niels Henrik I.Hjollund, Tina Kold Jensen, Tine Brink Henriksen, Henrik A.Kolstad, Erik Ernst, Anne-Marie Andersson, Aleksander Giwercman, Niels Erik Skakkebæk and Jørn Olsen.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Barratt, C.L.R., Tomlinson, M.J. and Cooke, I.D. (1993) Prognostic significance of computerized motility analysis for in vivo fertility. Fertil. Steril., 60, 520525.[ISI][Medline]
Bonde, J.P., Ernst, E., Jensen, T.K. et al. (1998a) The relation between semen quality and fertility. A population based follow-up study of 430 first-pregnancy planners. Lancet, 352, 11721177.[ISI][Medline]
Bonde, J.P., Hjollund, H., Jensen, T.K et al. (1998b) A follow-up study of environmental and biologic determinants of fertility among 430 Danish first-pregnancy planners: design and methods. Reprod. Toxicol., 12, 1927.[ISI][Medline]
ESHRE Andrology Special Interest Group (1996) Consensus workshop on advanced diagnostic andrology techniques. Hum. Reprod., 11, 14631479.
ESHRE Andrology Special Interest Group (1998) Guidelines on the application of CASA technology in the analysis of spermatozoa. Hum. Reprod., 13, 142145.
Irvine, D.S., Macleod, I.C., Templeton, A.A. et al. (1994) A prospective clinical study of the relationship between the computer-assisted assessment of human semen quality and the achievement of pregnancy in vivo. Hum. Reprod., 9, 23242334.[Abstract]
Jørgensen, N., Auger, J., Giwercman, A. et al. (1997) Semen analysis perfomed by different laboratory teams: an intervariation study. Int. J. Androl., 20, 201208.[ISI][Medline]
Krause, W. (1995) Computer-assisted semen analysis systems: comparison with routine evaluation and prognostic value in male fertility and assisted reproduction. Hum. Reprod., 10, 6066.
Lenzi, A. (1997) News from the European Academy of Andrology (EAA). Computer-aided semen analysis (CASA) 10 years later: a test-bed for the European scientific andrological community. Int. J. Androl., 20, 12.
Mortimer, D. (1994) Practical Laboratory Andrology. Oxford University Press, New York.
Mortimer, S.T. and Swan, M.A. (1999) Effect of image sampling frequency on established and smoothing-independent kinematic values of capacitating human spermatozoa. Hum. Reprod., 14, 9971004.
Paston, M.J., Sarkar, S., Oates, R.P. et al. (1994) Computer-aided semen analysis variables as predictors of male fertility potential. Archs Androl., 33, 9399.
World Health Organization (1992) WHO Laboratory Manual for the Examination of Human Semen and SpermCervical Mucus Interaction. Cambridge University Press, Cambridge.
Submitted on February 23, 1999; accepted on April 7, 2000.