ANDROLOGIA
24, 125- 129 ( 1992)
ACCEPTED: FEBRUARY 25, 1992
The sperm penetration test (P-test)can predict fecundability in the male partner from infertile couples E. Bostofte, P. Bagger, A. Michael and G. Stakemann Key words. Semen - P-test - male infertility - fertility prognosis - human.
Summary. Three hundred and twenty-one consecutive couples were investigated for infertility at Hvidovre University Hospital in the period from November 1977 to June 1985. The male partners were evaluated in two ways: the classical semen analysis, and the ability of sperm to penetrate fresh hen egg white, the P-test. A Cox regression analysis was used to describe the relation between these variables and fecundability, i.e. the time required to conceive. Four of thirteen variables- the number of morphologically normal spermatozoa, the number of motile spermatozoa, the P-test, and the man’s ageeach have significant relation to the fecundability. However, when covariation is considered, only the P-test and the man’s age posses significant prognostic information, whereas the variables of the classical semen analysis do not. This indicates that the P-test may replace the classical semen analysis when trying to predict individual pregnancy probabilities. Finally, the P-test and the man’s age are combined to form a prognostic index which predicts the fecundability of the male partner in the individual infertile couple. Introduction The classical semen analysis (WHO, 1987) predicts the chance for individual males from infertile couples to achieve pregnancy (Bostofte, 1987; Bostofte et al., 1990; Kjaergaard et al., 1990). The Sperm Penetration test (P-test), originally designed by Kremer (1965) and later standardized by the W H O (1987), has also been found to be correlated to male fertility (Ulstein, Department of Obstetrics and Gynecology, Hvidovre University Hospital, Copenhagen, Denmark. Correspondence: Dr Erik Bostofte, Bredevej 47B, 2830 Virum, Denmark.
1972; Insler et al., 1979; Alexander, 1981; Menge et al., 1982; Eggert-Kruse et al., 1989). The aim of the present investigation is to compare the values of the P-test and the classical semen analysis for predicting the achievement of pregnancy by men from infertile couples, without knowing the fertility of the women.
Material and methods Material The material consists of semen analysis from 321 consecutive men who, in the period November 30, 1977 to June 1, 1985, were examined at the Sperm Analysis Laboratory, Hvidovre University Hospital. The indication for semen analysis was infertility, i.e. they have tried to achieve pregnancy for a t least one year, and both partners had a complete fertility investigation, including ovulation detection, cervical mucus investigation, and anatomical conditions for the female, and semen analysis for the male. Only the first semen analysis from each male is included in the study, and the results of the female fertility investigation is not included. The period of observation starts on the date of the first semen analysis. The men were followed either until pregnancy was achieved or until discharge from the department after completion of fertility investigations or treatment. The time required to achieve conception is the period between the date of the first semen analysis and the date of the last menstrual period. Pregnancy was achieved by 108 men (34%), while 213 (66%) were observed for a n average of 2.3 years without having achieved pregnancy; their periods of observation serve as censored data.
Semen analysis All semen specimens were delivered within 30 min of ejaculation in a standard tube. The time from
126
E. BOSTOFTE ET AL.
ejaculation to semen analysis and the period of abstinence were registered. T h e appearance of the semen was described as normal, partly transparent, or transparent. T h e consistency of the semen was described as normal, partly viscous or viscous. Biochemical analysis and tests for sperm antibodies were not included in the semen analysis.
on a slide which was graded with 0.1 cm accuracy from 0 cm to 5.5 cm. A reservoir for the semen specimen was placed at the zero end, and the other end of the capillary was sealed with wax. The whole set up was placed horizontally in an incubator at 37 "C for 1 h. Thereafter, the distance of penetration for the fastest (sentinel) spermatozoon was measured.
Semen volume. Semen volume was measured in a graded tube with 0.1 ml accuracy.
Statistical methods
Spermatozoa motility and numbers of motile and immotile spermatozoa. These were determined not more than 2 h after ejaculation. One drop of semen was placed on a slide and covered with a cover slip. The numbers of immotile and motile spermatozoa expressed as percentage of the total number of spermatozoa were counted at ~ 6 0 0linear magnification using a square-field ocular (100-200 spermatozoa per count). Moreover, the number of white blood cells (none, few, some, or many) and red blood cells (present or not present) were determined. Spermatozoa motility was evaluated immediately and 3 h after receiving the semen sample using the same procedure and expressed as excellent, good, impaired, or poor (Hammen, 1944). Sperm count. Sperm count (million ml-I) was determined as follows: After mixing, 0.1 ml of semen was added to 1.9 ml of a 0.035% methylene blue solution, and the total number of spermatozoa was counted using a Biirker-Turk counting chamber. Numbers of morphologically abnormal spermatozoa. This was calculated as a percentage of total number of spermatozoa as follows: after mixing, semen was smeared on a glass slide. T h e smear was fixed in 96% methyl alcohol for 5 min followed by 70% methyl alcohol for 3 min. The smear was stained with Thymol blue and Janus green B (Michael et al., 1987, developed in 1975). If possible, no less than 200 spermatozoa were classified. Viability of spermatozoa ( h ) . Approximately 0.5 ml of mixed semen was placed in a dwarf tube, covered with a layer of mineral oil, and stored at room, temperature. A sample was withdrawn twice a day and examined microscopically as long as living spermatozoa were present. No staining method was used. The sperm penetration test (P-test) . A capillary tube filled with fresh hen egg white not more than 2-3 d old and free of air bubbles was placed
Cox regression analysis (Cox, 1972; Lee, 1980; Bostofte et al., 1990) were used to establish the relations between the variables of the semen analysis and the time required to conceive. The probability of conception before time (t) (P,) is calculated as follows:
P, = 1 -exp [exp (PI x Z, + - - -+fin x Z,) x In ( 1 -Pot)],
(1)
where Pot is the underlying probability of pregnancy before t. PI, - - , p, are the regression coefficients, and Z,, -, Z, are the values of the individual prognostic variables. The product of the regression coefficient and the value of the prognostic variable ( p x Z) is called the risk factor, and the exponential function of this [exp (p x Z)] is the risk factor index. This risk factor index has to be multiplied to the underlying pregnancy rate (equation 1) to get the actual pregnancy probability of the individual male. The model assumptions, i.e. proportionality of pregnancy rates, have been checked graphically by comparing plots of estimated integrated underlying pregnancy rates for relevant subgroups. Each continuous variable has been tested to determine whether a linear scoring is suitable or whether it should be replaced by some polynomial expression. A quotient test has been used. Also, in preliminary analyses, each continuous variable has been categorized in order not to overlook specific non-linear effects. ~
~
~
Variables The following variables were considered in order to discover their influence on the fecundability, i.e. the time required to conceive: (1) the man's age at semen analysis (years), range 19-73 years, median value 30 years; (2) presence of previous pregnancy (yeslno), 152 and 169 respectively; (3) semen volume (ml), range 0.3-9.9 ml, median value 3.0 ml (4) sperm count (million ml-I), range 0.04-284 million ml-I, median value 67 million ml-I; (5) number of morphological normal spermatozoa ( yo),range 15-82y0, median value 64%; (6) number of motile spermatoANDROLOGIA 24, 125-129 (1992)
MALEFERTILITY
zoa (yo),range 2-89%, median value 70%; (7) degree of motility (arbitrary units), 285 good, 36 bad; (8) the appearance of semen (arbitrary units); (9) the consistency of semen (arbitrary units); (10) the viability (h), 288 more than 48 h, 33 less than 48 h; (11) number of double forms (yo); (12) the P-test (cm), range 0.0-5.5 cm, median value 5.0 cm; ( 13) leucocytes in semen (arbitrary units). Results Table 1 shows the P values for the above mentioned 13 variables with possible significant prognostic information about fecundability during the foreward selection procedure of the Cox regression analysis. Three variables, i.e. number of morphologically normal spermatozoa, number of motile spermatozoa, and the P-test, each have individual significant relation to the time required to conceive. The man’s age at semen analysis has a P value of 0.0586. All the other variables are non-significant with higher P values. However, when co-variation is taken into consideration, number of morphologically normal spermatozoa and number of motile spermatozoa become insignificant, whereas the age of the man becomes significant (P
24, 125- 129 ( 1992)
ACCEPTED: FEBRUARY 25, 1992
The sperm penetration test (P-test)can predict fecundability in the male partner from infertile couples E. Bostofte, P. Bagger, A. Michael and G. Stakemann Key words. Semen - P-test - male infertility - fertility prognosis - human.
Summary. Three hundred and twenty-one consecutive couples were investigated for infertility at Hvidovre University Hospital in the period from November 1977 to June 1985. The male partners were evaluated in two ways: the classical semen analysis, and the ability of sperm to penetrate fresh hen egg white, the P-test. A Cox regression analysis was used to describe the relation between these variables and fecundability, i.e. the time required to conceive. Four of thirteen variables- the number of morphologically normal spermatozoa, the number of motile spermatozoa, the P-test, and the man’s ageeach have significant relation to the fecundability. However, when covariation is considered, only the P-test and the man’s age posses significant prognostic information, whereas the variables of the classical semen analysis do not. This indicates that the P-test may replace the classical semen analysis when trying to predict individual pregnancy probabilities. Finally, the P-test and the man’s age are combined to form a prognostic index which predicts the fecundability of the male partner in the individual infertile couple. Introduction The classical semen analysis (WHO, 1987) predicts the chance for individual males from infertile couples to achieve pregnancy (Bostofte, 1987; Bostofte et al., 1990; Kjaergaard et al., 1990). The Sperm Penetration test (P-test), originally designed by Kremer (1965) and later standardized by the W H O (1987), has also been found to be correlated to male fertility (Ulstein, Department of Obstetrics and Gynecology, Hvidovre University Hospital, Copenhagen, Denmark. Correspondence: Dr Erik Bostofte, Bredevej 47B, 2830 Virum, Denmark.
1972; Insler et al., 1979; Alexander, 1981; Menge et al., 1982; Eggert-Kruse et al., 1989). The aim of the present investigation is to compare the values of the P-test and the classical semen analysis for predicting the achievement of pregnancy by men from infertile couples, without knowing the fertility of the women.
Material and methods Material The material consists of semen analysis from 321 consecutive men who, in the period November 30, 1977 to June 1, 1985, were examined at the Sperm Analysis Laboratory, Hvidovre University Hospital. The indication for semen analysis was infertility, i.e. they have tried to achieve pregnancy for a t least one year, and both partners had a complete fertility investigation, including ovulation detection, cervical mucus investigation, and anatomical conditions for the female, and semen analysis for the male. Only the first semen analysis from each male is included in the study, and the results of the female fertility investigation is not included. The period of observation starts on the date of the first semen analysis. The men were followed either until pregnancy was achieved or until discharge from the department after completion of fertility investigations or treatment. The time required to achieve conception is the period between the date of the first semen analysis and the date of the last menstrual period. Pregnancy was achieved by 108 men (34%), while 213 (66%) were observed for a n average of 2.3 years without having achieved pregnancy; their periods of observation serve as censored data.
Semen analysis All semen specimens were delivered within 30 min of ejaculation in a standard tube. The time from
126
E. BOSTOFTE ET AL.
ejaculation to semen analysis and the period of abstinence were registered. T h e appearance of the semen was described as normal, partly transparent, or transparent. T h e consistency of the semen was described as normal, partly viscous or viscous. Biochemical analysis and tests for sperm antibodies were not included in the semen analysis.
on a slide which was graded with 0.1 cm accuracy from 0 cm to 5.5 cm. A reservoir for the semen specimen was placed at the zero end, and the other end of the capillary was sealed with wax. The whole set up was placed horizontally in an incubator at 37 "C for 1 h. Thereafter, the distance of penetration for the fastest (sentinel) spermatozoon was measured.
Semen volume. Semen volume was measured in a graded tube with 0.1 ml accuracy.
Statistical methods
Spermatozoa motility and numbers of motile and immotile spermatozoa. These were determined not more than 2 h after ejaculation. One drop of semen was placed on a slide and covered with a cover slip. The numbers of immotile and motile spermatozoa expressed as percentage of the total number of spermatozoa were counted at ~ 6 0 0linear magnification using a square-field ocular (100-200 spermatozoa per count). Moreover, the number of white blood cells (none, few, some, or many) and red blood cells (present or not present) were determined. Spermatozoa motility was evaluated immediately and 3 h after receiving the semen sample using the same procedure and expressed as excellent, good, impaired, or poor (Hammen, 1944). Sperm count. Sperm count (million ml-I) was determined as follows: After mixing, 0.1 ml of semen was added to 1.9 ml of a 0.035% methylene blue solution, and the total number of spermatozoa was counted using a Biirker-Turk counting chamber. Numbers of morphologically abnormal spermatozoa. This was calculated as a percentage of total number of spermatozoa as follows: after mixing, semen was smeared on a glass slide. T h e smear was fixed in 96% methyl alcohol for 5 min followed by 70% methyl alcohol for 3 min. The smear was stained with Thymol blue and Janus green B (Michael et al., 1987, developed in 1975). If possible, no less than 200 spermatozoa were classified. Viability of spermatozoa ( h ) . Approximately 0.5 ml of mixed semen was placed in a dwarf tube, covered with a layer of mineral oil, and stored at room, temperature. A sample was withdrawn twice a day and examined microscopically as long as living spermatozoa were present. No staining method was used. The sperm penetration test (P-test) . A capillary tube filled with fresh hen egg white not more than 2-3 d old and free of air bubbles was placed
Cox regression analysis (Cox, 1972; Lee, 1980; Bostofte et al., 1990) were used to establish the relations between the variables of the semen analysis and the time required to conceive. The probability of conception before time (t) (P,) is calculated as follows:
P, = 1 -exp [exp (PI x Z, + - - -+fin x Z,) x In ( 1 -Pot)],
(1)
where Pot is the underlying probability of pregnancy before t. PI, - - , p, are the regression coefficients, and Z,, -, Z, are the values of the individual prognostic variables. The product of the regression coefficient and the value of the prognostic variable ( p x Z) is called the risk factor, and the exponential function of this [exp (p x Z)] is the risk factor index. This risk factor index has to be multiplied to the underlying pregnancy rate (equation 1) to get the actual pregnancy probability of the individual male. The model assumptions, i.e. proportionality of pregnancy rates, have been checked graphically by comparing plots of estimated integrated underlying pregnancy rates for relevant subgroups. Each continuous variable has been tested to determine whether a linear scoring is suitable or whether it should be replaced by some polynomial expression. A quotient test has been used. Also, in preliminary analyses, each continuous variable has been categorized in order not to overlook specific non-linear effects. ~
~
~
Variables The following variables were considered in order to discover their influence on the fecundability, i.e. the time required to conceive: (1) the man's age at semen analysis (years), range 19-73 years, median value 30 years; (2) presence of previous pregnancy (yeslno), 152 and 169 respectively; (3) semen volume (ml), range 0.3-9.9 ml, median value 3.0 ml (4) sperm count (million ml-I), range 0.04-284 million ml-I, median value 67 million ml-I; (5) number of morphological normal spermatozoa ( yo),range 15-82y0, median value 64%; (6) number of motile spermatoANDROLOGIA 24, 125-129 (1992)
MALEFERTILITY
zoa (yo),range 2-89%, median value 70%; (7) degree of motility (arbitrary units), 285 good, 36 bad; (8) the appearance of semen (arbitrary units); (9) the consistency of semen (arbitrary units); (10) the viability (h), 288 more than 48 h, 33 less than 48 h; (11) number of double forms (yo); (12) the P-test (cm), range 0.0-5.5 cm, median value 5.0 cm; ( 13) leucocytes in semen (arbitrary units). Results Table 1 shows the P values for the above mentioned 13 variables with possible significant prognostic information about fecundability during the foreward selection procedure of the Cox regression analysis. Three variables, i.e. number of morphologically normal spermatozoa, number of motile spermatozoa, and the P-test, each have individual significant relation to the time required to conceive. The man’s age at semen analysis has a P value of 0.0586. All the other variables are non-significant with higher P values. However, when co-variation is taken into consideration, number of morphologically normal spermatozoa and number of motile spermatozoa become insignificant, whereas the age of the man becomes significant (P
