Prediction of pregnancy by intrauterine insemination using CASA estimates and strict criteria in patients with male factor infertility

Prediction of pregnancy by intrauterine insemination using CASA estimates and strict criteria in patients with male factor infertilityCASA estimates and strict criteria in patients with male
factor infertility
YUKI HIRANO, TATSUYA SUZUKI, AKIKO OHNO, SATORU
TAKAMIZAWA and MITSUAKI SUZUKI
Department of Obstetrics and Gynecology, Jichi Medical School, Tochigi, Japan
Summary This study was performed to predict pregnancy by intrauterine insemination (IUI)
using computer-aided sperm analysis (CASA) estimates and strict criteria in patients with
male factor infertility. IUI was performed in 682 cycles on 160 male factor infertile
patients. Semen examinations were carried out by CASA and strict criteria before and
after sperm preparation using continuous-step density gradient centrifugation. Receiver
operating characteristics (ROC) curves were constructed for assessment of the
effectiveness of each individual parameter in predicting pregnancy by IUI. A clinically
acceptable threshold was calculated when sensitivity plus specificity were maximum. The
average cycle of IUI performed was 4.3 ± 2.4. Pregnancy rate per cycle and per patient
were 7.2% (49/682) and 28.1% (45/160), respectively. Using ROC curve, it was shown
that normal sperm morphology assessed by the strict criteria before sperm separation and
five parameters after sperm separation including rapid, progressive motility, average path
velocity (VAP), curvilinear velocity (VCL), and straight line velocity (VSL) were able to
predict pregnancy by IUI. Correlation between sperm parameters and pregnancy
outcome was examined by the logistic regression model. In a multivariate analysis normal
morphology before sperm separation ‡15.5% [odds ratio (OR) ¼ 2.2, p ¼ 0.02], rapid
after sperm separation ‡25.5% [OR ¼ 3.9, p ¼ 0.029], and VCL after sperm separation
‡102.65 lm/sec [OR ¼ 3.2, p ¼ 0.002] were the parameters of predictive value for
pregnancy outcome. Adjustment of the model for female age, female infertility factors,
and the methods of ovulation induction did not change this finding, and the final model
still had the same covariates. Pregnancy rates per cycle according to the number (0, 1, 2
and 3) of variables satisfied with the three parameters were 0% (0/110), 1.6% (3/183),
9.7% (21/217) and 15.1% (23/151), respectively. Three semen parameters including
normal morphology before sperm separation, rapid and VCL after sperm separation were
identified as predictors of pregnancy by IUI. These variables would be helpful when
counselling patients before they make the decision to proceed with in vitro fertilization
(IVF)/intracytoplasmic sperm injection (ICSI)-ET.
pregnancy, strict criteria
machi, Kawachi-gun, Tochigi 329-0498, Japan. E-mail: [email protected]
international journal of andrology, 27:63–68 (2004)
2004 Blackwell Publishing Ltd.
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
ductive technologies (ARTs) including in vitro fertilization
(IVF) and intracytoplasmic sperm injection (ICSI) have been
developed, artificial intrauterine insemination (IUI) with
husbands semen has been playing an important role in the
infertility treatment. Indications for IUI have included
oligozoospermia, asthenozoospermia, pyospermia, hostile
Because of the diversity of its indications, variable degrees of
success in IUI depend on the type of male or female problem
for which the procedure was used.
Our group has been investigating sperm fertilizing ability
using several sperm parameters and sperm function tests,
including computer-aided sperm analysis (CASA) (Hirano
et al., 2001), normal sperm morphology using the strict
criteria (Obara et al., 2001), the zona-free hamster egg
penetration test (Shibahara et al., 1998), and Sperm Quality
Analyzer (SQA) (Shibahara et al., 1997, 1999). Moreover,
we have utilized the hemizona assay (HZA) to demonstrate
the inhibitory effects of antibodies against sperm (Shibahara
et al., 1991, 1993, 1996) or zona-pellucida (Afzalpurkar
et al., 1997) on fertilization. However, we have not yet
investigated which sperm parameter has the predictive value
of the success in IUI treatments. Marshburn et al. (1992)
reported that some of the CASA estimates after thawing and
washing of cryopreserved sperm could be the predictors of
fertility outcome after IUI. A recent review by Van Waart
et al. (2001) has shown that the predictive value of normal
sperm morphology in IUI has still been controversial. These
results may indicate that there is a limitation for predicting
fertility outcome after IUI using a single sperm parameter.
This study was performed to predict pregnancy after IUI
using the CASA estimates and the strict criteria in patients
with male factor infertility.
Materials and methods
Subjects A total of 160 couples with male factor at the hospital in
Jichi Medical School were enrolled into this study and their
characteristics are summarized in Table 1. Definition of
patients with male factor infertility is based on the
examinations of ejaculated semen carried out at least twice
as assessed by the criteria of World Health Organization
(1999). The diagnosis for 160 patients with male factor
examined by CASA and strict criteria included 94 terato-
asthenozoospermia, 46 asthenozoospermia, nine oligo-
astheno-teratozoospermia, nine oligozoospermia, and two
oligo-asthenozoospermia. IUI was performed in 682 cycles
in the study period between May 1995 and December 1999.
Infertile women aged older than 40, those with sperm
immobilizing antibodies in their sera, and those with
congenital uterine abnormalities were excluded from this
study. Two hundred and thirty-six cycles were treated in the
spontaneous ovulation cycles, while 446 cycles were treated
with the use of ovarian stimulants including bromocriptine
or clomiphene citrate or gonadotrophin. The mean age of
the male and female subjects were 34 ± 5 years (range 24–
52) and 31 ± 4 years (23–39), respectively. The average
infertility period was 2.8 ± 2.0 years (range 0.1–11.0). One
hundred and twenty of 160 couples were diagnosed primary
sterility. Semen examinations were carried out using CASA
and strict criteria before and after sperm preparation as
described below. Clinical pregnancy was diagnosed when
the gestational sac was demonstrated under transvaginal
ultrasonography.
Semen analysis using CASA and strict criteria After liquefaction, semen quality analysis was performed
using the CASA system (Hamilton Thorne Research,
Beverly, MA, USA) for the 682 IUI treatment cycles.
Briefly, a 5-lL aliquot of semen sample was placed on a
Makler chamber. At least 200 sperm were counted with
CASA to evaluate the sperm concentration, sperm motility,
and sperm motion variables including rapid sperm movement
(rapid), curvilinear velocity (VCL), straight line velocity
(VSL), average path velocity (VAP), amplitude of lateral head
displacement (ALH), linearity (LIN ¼ VSL/VCL), and
straightness (STR ¼ VSL/VAP). The CASA settings were
followed according to the manufacturer’s instruction.
The sperm morphology was assessed following the
method by Kruger et al. (1986, 1988). Briefly, 5 lL semen
were used to make the smears using two morphology slides
cleaned thoroughly with 70% ethanol before use. The slides
were air-dried at room temperature and fixed for 15 sec with
Diff-Quik fixative (Diff-Quik, International Reagents Cor-
poration, Kobe, Japan) prior to staining with Diff-Quik
solutions. The slides were read and documented on the same
day. Spermatozoa were considered as normal when the head
had a smooth oval configuration with a well-defined
Table 1. Characteristics of 160 patients with male factor treated by intrauterine insemination (IUI)
Mean ± SD Range
Age Husband (years) 34 ± 5 24–52 Wifea (years) 31 ± 4 23–39
Infertility period (years) 2.8 ± 2.0 0.1–11 Infertility
Primary (cases) 120 Secondary (cases) 40
No. of IUI performed (times) 4.3 ± 2.4 1–12 PR/cycle (%) 7.2 (49/682) PR/patient (%) 28.1 (45/160)
aWomen P 40 years uterine anomaly and immunological infertility were excluded; PR, pregnancy rate.
64 H. Shibahara et al.
2004 Blackwell Publishing Ltd, International Journal of Andrology, 27, 63–68
13652605, 2004, 2, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
acrosome involving about 40–70% of the sperm head, as well
as an absence of neck, midpiece, or tail defects. No
cytoplasmic droplets of more than half the size of the sperm
head were allowed to be present. The length of a normal
sperm head was 5–6 lm and the diameter 2.5–3.5 lm. At
least 200 cells per slide were evaluated.
Procedure of sperm preparation The selective concentration of progressively motile sperm
for IUI with husbands semen was carried out using a
modified method of continuous-step density gradient cen-
trifugation demonstrated by Kaneko et al. (1987). Briefly,
the ejaculated semen was placed on 6.0 mL of 80% Percoll
solution, and the density gradient was formed by mixing the
semen and Percoll using an L-shaped rod for two or three
strokes. After centrifugation at 600 g for 20 min, progres-
sively motile sperm were concentrated in the sediment,
whereas the immotile sperm and other types of cells
remained in the upper part of the density gradient. The
supernatant was removed and the pellet was resuspended in
0.5 mL of Hanks medium. Two 5-lL aliquots were used for
CASA and to make the smear for the morphological
assessment. The rest of the suspension was used for IUI.
Statistical methods A univariate logistic regression model was used to test the
above variables for their association with fertility outcome
after IUI. Variables that were statistically significant associ-
ated with establishing pregnancy were retained for testing
in a multivariate logistic regression model (Hosmer &
Lemeshow, 1989). Receiver operating characteristics
(ROC) curves (Hanley & McNeil, 1982; Zweig &
Campbell, 1993) were constructed for assessment of the
effectiveness of each individual parameter in predicting
pregnancy by IUI. A clinically acceptable threshold was
calculated when sensitivity plus specificity was maximum. In
the multivariate analyses the results were adjusted for the
possible confounding factors. These factors include female
age, infertility factors in women, the method of ovulation
induction, and bromocriptin administration.
The statistical analyses were carried out using the
Statistical Package for Social Science (SPSS) for personal
computers, version 10.0 for Windows (SPSS Institute Inc.,
Chicago, IL, USA). A value of p < 0.05 was defined as
representing a significant difference.
Results
Results of IUI treatment As shown in Table 1, the average numbers of cycle
treated by IUI were 4.3 ± 2.4 (range 1–12). Forty-five of
160 infertile patients established pregnancy, giving a preg-
nancy rate per patient of 28.1%. Four patients conceived
twice by IUI in the study period. The pregnancy rate per
cycle was 7.2% (49/682).
Prediction of fertility outcome after IUI ROC curves were constructed for assessment of the
effectiveness of each individual parameter in predicting
pregnancy by IUI (Fig. 1). A clinically acceptable threshold
was calculated when sensitivity plus specificity were
maximum. Using ROC curve, it was shown that normal
sperm morphology assessed by the strict criteria before
sperm preparation and five parameters after sperm prepar-
ation including % rapid, progressive motility, VAP, VCL,
and VSL were able to predict pregnancy by IUI (Table 2).
Correlation between sperm parameters and fertility out-
come after IUI was examined using the logistic regression
model. In a multivariate analysis, normal morphology
before sperm separation ‡15.5% [odds ratio (OR) ¼ 2.2,
p ¼ 0.02], % rapid after sperm separation ‡25.5%
Table 2. Construction of the threshold to predict pregnancy after intrauterine insemination (IUI) using ROC curves
Parameter Thresholda
Before preparation Normal morphology (%) 15.5
After preparation Rapid (%) 25.5 Progressive motility (%) 25.5 VAP (lm/sec) 63.65 VCL (lm/sec) 102.65 VSL (lm/sec) 52.25
aA clinically acceptable threshold was calculated when sensitivity plus specificity was maximum.
ROC curve
1-Specificity 1.000.750.500.25
S en
si tiv
Normal morphology
% rapid (post)
% progressive (post)
VAP (post)
VCL (post)
VSL (post)
Figure 1. Construction of the ROC curves. The ROC curves were construc- ted for assessment of the effectiveness of each individual parameter in predicting pregnancy by intrauterine pregnancy (IUI). VAP, average path velocity; VCL, curvilinear velocity; VSL, straight line velocity.
Prediction of pregnancy by IUI in male factor infertility 65
2004 Blackwell Publishing Ltd, International Journal of Andrology, 27, 63–68
13652605, 2004, 2, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
[OR ¼ 3.9, p ¼ 0.029], and VCL after sperm separation
‡102.65 lm/sec [OR ¼ 3.2, p ¼ 0.002] were the param-
eters of predictive value for pregnancy outcome (Table 3).
Adjustment of the model for female age, infertility factors in
women, methods of ovulation induction and bromocriptine
administration did not change this finding, and the final
model still had the same covariates. Pregnancy rates per
cycle according to the number (0, 1, 2 and 3) of variables
satisfied with the three parameters were 0% (0/110), 1.6%
(3/183), 9.7% (21/217) and 15.1% (23/151), respectively
(Fig. 2). There were significant differences of the pregnancy
rate per cycle between 0 and 2 variables exceeding the
threshold (p < 0.001), and between 0 and 3 variables
exceeding the threshold (p < 0.001). There were also
significant differences of the pregnancy rate per cycle
between 1 and 2 variables exceeding the threshold
(p < 0.001), and between 1 and 3 variables exceeding the
threshold (p < 0.001).
Discussion For couples with male factor infertility, the least invasive
procedure in ARTs is IUI. IUI should generally be used as a
first-line treatment in male factor infertility after treatment
of specific factors affecting male fertility but prior to
proceeding with other more invasive ARTs. In such
infertile patients, the success of IUI is dependent on semen
quality. A critical review for IUI by Allen et al. (1985)
demonstrated that in the results of 18 series from 1957 to
1984, a total of 714 patients were inseminated, with an
overall conception rate per patient of 28% (range 0–62%).
In the article, only traditional semen parameters including
sperm concentration, sperm motility, sperm morphology,
and zona-free hamaster egg penetration rate were analysed
to predict pregnancies after IUI in infertile couples with
male factor. Later, CASA systems that can identify and track
human sperm have been developed to revolutionize the
research of the movement of human sperm (Aitken et al.,
1985; Jeulin et al., 1986; Katz et al., 1986). The CASA has
the advantages of providing objective semen analysis data on
sperm kinetics. Our group has been using CASA to predict
sperm fertilizing ability in IVF (Hirano et al., 2001),
comparing its functions with other devices such as SQA
(Suzuki et al., 2002), and analyse the association with
multinucleate formation in IVF (Shibahara et al., 2003).
There is a study that utilized CASA to investigate the
relationship between sperm characteristics from fresh and
frozen donor semen with fertility outcome after IUI
(Marshburn et al., 1992). They concluded that some of
the sperm motion characteristics using CASA after thawing
and washing of cryopreserved sperm are the predictors of
fertility outcome after IUI.
IUI, Van Waart et al. (2001) recently summarized the
contents of 18 papers demonstrating the predictive value of
normal sperm morphology in IUI. Of the nine publications
that used the strict criteria for sperm morphology (Kruger
et al., 1986, 1988), six stated a positive predictive value for
sperm morphology (Irianni et al., 1993; Toner et al., 1995;
Lindheim et al., 1996; Ombelet et al., 1996, 1997;
Montanaro Gauci et al., 2001). However, other reports
could not show the correlation between normal sperm
morphology and the pregnancy outcome after IUI (Matorras
et al., 1995; Karabinus & Gelety, 1997; Shulman et al.,
1998). Thus, there seems to be a limitation for predicting
fertility outcome after IUI using a single sperm parameter.
In the present study, a retrospective, univariate, multiva-
riate and ROC analysis of a large series of 682 consecutive
IUI in a single centre were carried out. As shown in Table 2,
it was shown that normal sperm morphology assessed by the
strict criteria before sperm separation and five parameters
after sperm separation including rapid, progressive motility,
VAP, VCL, and VSL were able to predict pregnancy by IUI.
In them, it was shown after an adjustment for possible
confounding factors that normal morphology before sperm
Table 3. Odds ratio (OR) adjusted for possible confounding factorsa in the multivariate analyses
B p OR 95% Cl of OR
Before preparation Normal morphology (P15.5%)
0.785 0.029 2.192 1.085–4.429
After preparation % Rapid (P25.5%) 1.449 0.030 4.257 1.150–15.700 VCL (P102.65 lm/sec)
1.363 <0.001 3.906 1.527–8.384
aThe model was adjusted for the following factors: female age, infertile factors in women, method of ovulation induction and bromocriptine admin- istration; B, estimate; Cl, confidence interval.
0
2
4
6
8
10
12
14
16
n = 183
n = 217
n = 151
P er
ce nt
ag e
Figure 2. Association between pregnancy rate and number of variables exceeding the threshold in the three independent predictive parameters. Pregnancy rates per cycle according to the number (0, 1, 2 and 3) of variables satisfied with the three parameters were 0% (0/110), 1.6% (3/183), 9.7% (21/217) and 15.1% (23/151), respectively, *,**,***,****p < 0.001.
66 H. Shibahara et al.
2004 Blackwell Publishing Ltd, International Journal of Andrology, 27, 63–68
13652605, 2004, 2, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
separation ‡15.5% [OR ¼ 2.2, p ¼ 0.02], rapid after sperm
separation ‡25.5% [OR ¼ 3.9, p ¼ 0.029], and VCL after
sperm separation ‡102.65 lm/sec [OR ¼ 3.2, p ¼ 0.002]
were the parameters of predictive value for pregnancy
outcome (Table 3). We have previously shown that VCL
and rapid before and after sperm swim-up provide reliable
estimation of the fertilization ability of human sperm (Hirano
et al., 2001). It was suggested in the report that the total
distance travelled by rapid sperm movement might be
important in human sperm fertilizing abilities. The present
study also supports the importance of these sperm motion
variables for sperm movement to reach the site where sperm
meet ovulated oocyte.
As shown in Fig. 2, pregnancy rates per cycle according
to the number (0, 1, 2 and 3) of variables satisfied with
the three parameters were 0% (0/110), 1.6% (3/183), 9.7%
(21/217) and 15.1% (23/151), respectively. So far, this is the
first identification of the predictors of pregnancy after IUI
using CASA estimates and strict criteria in patients with male
factor infertility.
morphology before sperm separation, rapid and VCL after
sperm separation were identified as predictors of pregnancy
by IUI. The information obtained might be useful in
planning strategies for further treatment of infertile couples
with male factor to proceed with IVF/ICSI-ET.
References Afzalpurkar, A., Shibahara, H., Hasegawa, A., Koyama, K. &
Gupta, S. K. (1997) Immunoreactivity and in-vitro effect on
human sperm-egg binding of antibodies against peptides cor-
responding to bonnet monkey zona pellucida-3 glycoprotein.
Human Reproduction 12, 2664–2670.
Aitken, R. J., Sutton, M., Warner, P. & Richardson, D. W. (1985)
Relationship between the movement characteristics of human
spermatozoa and their ability to penetrate cervical mucus and
zona-free hamster oocytes. Journal of Reproduction and Fertility 73,
441–449.
Allen, N. C., Herbert III, C. M., Maxson, W. S., Rogers, B. J.,
Diamond, M. P. & Wentz, A. C. (1985) Intrauterine in-
semination: a critical review. Fertility and Sterility 44, 569–580.
Hanley, J. A. & McNeil, B. J. (1982) The meaning and use of the
area under receiving operating characteristics (ROC) curve.
Radiology 143, 29–36.
S., Yamaguchi, C., Tsunoda, H. & Sato, I. (2001)
Relationships between sperm motility characteristics assessed
by the computer-aided sperm analysis (CASA) and fertilization
rates in vitro. Journal of Assisted Reproduction and Genetics 18,
213–218.
Regression. New York, NY: John Wiley & Sons.
Irianni, F. M., Ramey, J., Vaintraub, M. T., Oehninger, S. &
Acosta, A. A. (1993) Therapeutic intrauterine insemination
improves with gonadotrophin ovarian stimulation. Archives of
Andrology 31, 55–62.
Jeulin, C., Feneux, D., Serres, C., Jouannet, P., Guillet-Rosso, F.,…