FERTILITY AND STERILITY Vol. 49, No.

1, January 1988
Copyright e 1988 The American Fertility Society Printed in U.S.A.

Predictive value of abnormal sperm morphology in

in vitro fertilization

Thinus F. Kruger, M.D.* R. James Swanson, Ph.D.§

Anibal A. Acosta, M.D.t:\: James F. Matta, Ph.D.§
Kathryn F. Simmons, M.S.§ Sergio Oehninger, M.D.t

Tygerberg Hospital, University of Stellenbosch, Parow, South Africa, and Eastern Virginia Medical School, Norfolk, Virginia

In patients with acceptable sperm count and motility, two patterns of abnormal mor-
phology, judged with strict criteria, were identified and described. Patients with <4%
normal forms and <30% morphology index (summation of normal and slightly amorphous
forms) had a fertilization rate of 7.6% of the oocytes (P pattern, poor prognosis). Patients
with normal morphology between 4 and 14% had a significantly better fertilization rate of
63.9% of the oocytes (P < 0.0001). Cases with >14% normal forms fertilized within the
normal range for the laboratory. By evaluating sperm morphology with the proposed strict
criteria, its predictive value in in vitro fertilization is enhanced. Fertil Steril 49:112, 1988

Although there is still extensive debate about the Although there was severe impairment in the fer-
role of sperm morphology in in vitro fertilization tilization rate, some of these patients still fertilized
(IVF)/-3 the human model has greatly improved the human egg; in these cases, a pregnancy was
the understanding of the significance of this pa- possible. 4
rameter for fertilization and pregnancy outcome. 4,5 The purpose of this study was to evaluate pa-
In previous publications 1,4 it was noted that if eval- tients with normal sperm morphology < 14% to try
uation of normal sperm morphology is done using to establish a morphologic pattern which can dif-
strict criteria, this parameter has an excellent pre- ferentiate the subgroup that fertilized from the
dictive value of fertilization. In patients with a subgroup that did not.
sperm concentration> 20 X 106 /ml and a motility
of >30% with a normal sperm morphology of
MATERIALS AND METHODS
<14%, the fertilization rate was markedly impaired
(37% to 47% per oocyte), as opposed to a high fer- Forty-five couples were allocated to the study
tilization rate (85% to 88%) when normal morphol- group in a prospective way. All female partners in
ogy was >14%.1,4 these couples had tubal infertility, and the males
had either been considered normal or had some
abnormal parameters by other laboratories evalua-
tions. Twenty-eight patients were stimulated with
Received February 16, 1987; revised and accepted August 19, a combination of hFSH/hMG/hCG (human folli-
1987.
* Infertility Clinic, Tygerberg Hospital and University of cle-stimulating hormone/human menopausal go-
Stellenbosch. nadotropin/human chorionic gonadotropin; 62.2%),
t Department of Obstetrics and Gynecology, Jones Institute 13 with hFSH/hCG (28.8%), and 4 with hMG/hCG
for Reproductive Medicine, Eastern Virginia Medical School. (8.8%) following protocols previously published. 6
:j: Reprint requests: Anibal A. Acosta, M.D., Hofheimer Hall, In the Norfolk experience, all of these protocols
6th Floor, Howard and Georgeanna Jones Institute for Repro-
ductive Medicine, 825 Fairfax Avenue, Norfolk, Virginia 23507.
have demonstrated provision of preovulatory 00-
§ Andrology Laboratory, Department of Biological Sciences, cytes with identical fertilization rates. All male pa-
Old Dominion University and Eastern Virginia Medical School. tients had to have a sperm concentration > 20

112 Kruger et al. Abnormal sperm morphology and IVF Fertility and Sterility
X 106 /ml and a progressive motility of >30%4,7 in
the basic semen analysis to try to minimize the
impact of these two variables on the fertilization
rate. The basic semen evaluation was performed
after liquefaction of the specimen delivered for IVF
insemination using computer analysis (Cellsoft
Semen Analysis System, Labsoft Division of Cryo
Resources Ltd., NY). Sperm concentration and
percentage of normal motility were assessed in this
fashion. Two morphology slides were prepared for 3. 4.
each patient from the specimen delivered on the
day of laparoscopy for in vitro insemination after
liquefaction and were stained by the quick-stain
technique. 8 Special care was taken to clean the
slides thoroughly with 70% ethyl alcohol before
using them, and no more than 5 JLI of semen were
used in order to make the smears as thin as possi- Figure 1 Diagramatic representation of quick-stained sper-
ble. The slides were air-dried at room temperature, matozoa. a. Normal form; head, oval shape, smooth configura-
were fixed for 15 seconds with Diff-Quik fixative tion, acrosome 40% to 70%, no neck, midpiece, or tail defects.
(Diff-Quik AHS del Caribe, Inc. Aguada, PR 00602) Head length; 5 to 6 ",m, diameter 2.5 to 3.5 ",m. b. 1. Slightly
amorphous head; slightly elongated, loss of oval shape, acro-
(1.8 mg/l triarylmethane methyl alcohol) prior to some 40% to 70%, diameter 2 to 2.5 ",m. b. 2. Slightly amorphous
staining with Diff-Quik solution 1 (1 gm/l xanthene neck defect; thick neck but normal-shaped head. c. Severe
in sodium azide-preserved buffer) for 10 seconds, amorphous forms; abnormalities in shape and acrosome. c. 1, 2.
Abnormally small acrosome. c. 3. No acrosome. c. 4. Acrosome
and then with solution 2 (0.625 gm/l azure A and > 70% of head.
0.625 gmll methylene blue in buffer) for 5 seconds.
In between the fixing step and each of the staining JLm, with slight abnormalities in the shape of the
steps, the excess solutions were drained from the head, but with a normal acrosome (Fig. 1). Severe
slides by blotting the slide edges on bibulous paper. amorphous head abnormalities were defined as
The slides were read on the same day and docu- those with no acrosome at all or those with an
mented. The morphology was evaluated, as out- acrosome smaller than 30% or larger than 70% of
lined in detail by Kruger et al, 1 by two independent the sperm head (Fig. 1). Completely abnormal
observers, each unaware of the results obtained by shapes also were put into this category (Fig. 1).
the other. This method of evaluation has an inter- Neck defects were also classified into two catego-
technician coefficient of variation and an intra- ries: slightly amorphous and severely amorphous
technician variability that are not significant neck defects. The slight neck defects referred to
(Spearman's rank correlation coefficient, r those sperm with debris around the neck or a
= 0.8695 and 0.9650, respectively).8,9 thickened neck, but with a normally shaped head
Spermatozoa were considered normal when the (Fig. 1). The severe defects referred to those sperm
head had a smooth oval configuration with a well- with a bend in the neck or midpiece of more than
defined acrosome involving about 40% to 70% of 30% or a severely amorphous head shape, as de-
the sperm head, as well as an absence of neck, scribed. All other abnormal sperm forms-round,
midpiece, or tail defects. No cytoplasmic droplets small, large, tapered, double head, double or coiled
of more than half the size of the sperm head should tail, cytoplasmic droplets-were classified follow-
be present. 1 The length of a normal sperm head ing the World Health Organization classificationY
was 5 to 6 JLm and the diameter 2.5 to 3.5 JLm (Fig. Female or male patients with antisperm antibod-
1). A micrometer in the eyepiece ofthe microscope ies were excluded from this study.
was used to do the routine measurements. In con- The human IVF procedures used for sperm prep-
trast to other methods,10 borderline forms were aration, insemination, and culture in the Norfolk
counted as abnormal. At least 200 cells per slide program have been described previously.12 Only
were evaluated. The amorphous-head group was mature oocytes with an extruded polar body were
divided into two categories: slightly amorphous and used in this study; 50,000 to 100,000 sperm/ml/egg
severely amorphous. Slightly amorphous forms were used for oocyte insemination in a total of 3.0
were those sperm with a head diameter of 2.0 to 2.5 ml of insemination medium.

Vol. 49, No.1, January 1988 Kruger et al. Abnormal sperm morpfwlogy and IVF 113
 After completion of the study, the results of the Table 1 Abnormal Morphology as a Predictor
of Human IVF: Semen Analysis
hamster tests in patients that had the assay per-
formed at least 8 weeks before the IVF procedure Group 1 Group II
were evaluated. The hamster tesi; was performed as (n = 13) (n = 32)
outlined by Swanson et al. I3 Penetration > 20%
Normal forms (%) 1.8 (2.4) 7.7 (3.3)"
was considered good, between 11 and 19% doubtful, Slightly amorphous
and <10% poor. The donors used as control always (head and neck) 18.0 (10.9) 34.3 (6.7)"
penetrated above the 20% level. Morphology index (slightly
amorphous forms and
The percentage of normal morphology, the con- normal) 19.7 (11.7) 42.0 (7.8)"
centration, and motility were noted carefully in Concentration (million/ml) 63.3 (42.8) 83.3 (57.8)b
each case, as were the fertilization and cleavage Motility (%) Mean
(standard deviation) 45.6 (13.2) 55.3 (18.6)b
rates and pregnancy outcome. The relationships
between the sperm parameters and the fertilization "P < 0.0001.
rates were examined using multiple regression b Not significant.

analysis in the Statistical Analysis System (SAS)

general linear model (GLM) procedure. The SAS significant difference) (Table 1). By eliminating
GLM procedure allows examination of all sub- abnormal concentration and abnormal motility, we
models of the complete multiple regression model. tried to individualize and define the effect of mor-
The multiple regression analysis examines the phology and its abnormalities in the process of fer-
contribution of all the independent variables to the tilization. There was a significant difference be-
variation in the dependent variable fertilization. tween the percent of normal morphology (1.8
Fertilization was standardized for the number of ± 2.4% in group I and 7.7 ± 3.3% in group II; P
preovulatory eggs by dividing the number of eggs < .0001) and the percentage of slightly amorphous
exhibiting fertilization by the number of preovula- abnormalities (head and neck), which was 18.0
tory eggs. Pregnancy rate per laparoscopy was cal- ± 10.9% in group I and 34.3 ± 6.7% in group II (P
culated by dividing the total number of laparosco- < .0001; Table 2). None of the other variables
pies by the number of pregnancies in the study. showed a significant difference between the two
Pregnancy rate per embryo transfer was calculated groups.
by dividing the number of patients who reached the The predictive value of normal morphology (r2
transfer stage by the number of pregnancies. The = 0.44) was better than that of slightly amorphous
following variables were evaluated: percent normal forms (r2 = 0.36). When we added the percent of
sperm morphology, amorphous head abnormalities normal morphology and slightly amorphous abnor-
(slight and severe), neck abnormalities (slight and malities (morphology index) and performed re-
severe), small, large, round, tapered, and double gression analysis, there was a highly significant
heads, cytoplasmic droplets, and tail abnormalities correlation between that index and fertilization (P
(double and coiled). < .0001; Table 1), with an even better predictive
value (r2 = 0.56). The mean for morphology index
RESULTS was 19.7 ± 11.7% in group I and 42 ± 7.8% in group
Of the 45 patients included in this study, 13 II (Table 1).
(28.9%) did not fertilize any oocytes at all, 17 The SAS general linear model was used with the
(37.8%) fertilized <50% of the oocytes obtained, number of embryos as the dependent variable to
and 15 (33.3%) fertilized >50% of oocytes obtained. determine a threshold to indicate where the
This should be compared with a fertilization rate chances of fertilization were significantly impaired.
for patients with tubal infertility of 89% to 92% in A threshold of 4% was indicated for normal mor-
our laboratory. phology and 30% for the combination of normal
The patients were divided into two groups: group morphology and slightly amorphous forms (mor-
I, 14 patients (no fertilization) and group II, 32 phology index). The fertilization rate per oocyte in
patients (fertilization of at least one oocyte). The group I (morphology index <30%, normal morphol-
mean sperm concentration in group I was 63.3 ogy <4%) was 7.6% and in group II (morphology
± 42.8 X 106 /ml (mean ± standard deviation) and index >30%, normal morphology >4%) was 63.9%
in group II, 83.3 ± 57.8 X 106 /ml (no significant (Table 2).
difference) (Table 1). The mean motility in group I The mean number of embryos in the 13 patients
was 45.6 ± 13.2% and in group II, 55.3 ± 18.6% (no in group I was 0.4 and for the 32 patients in group II

114 Kruger et al. Abnormal sperm morphology and IVF Fertility and Sterility
Table 2 Abnormal Morphology as a Predictor suIts also indicate that by adding the slightly
of Human IVF: Fertilization Rate amorphous forms to the normal forms, a "mor-
45 Patients, cone. > 20 million, motility> 30%
phology index" can be established with a cutoff
figure at the 30% level. Patients with a value of
Group I Group II
(n = 13) (n = 32)
<30% morphology index will have a severe reduc-
tion in fertilization as compared with patients hav-
P pattern G pattern ing an index> 30% (P < 0.0001) (Table 2). None of
Fertilization rate (%) (per oocyte) 7.6 63.9" the other semen parameters evaluated were of any
Mean no. embryos (per patient) 0.4 2.6" help to predict a patient's chance to fertilize.
a p = 0.0001. The advantage of strict morphology evaluation is
the fact that it is reproducible between patients and
was 2.6; these means were significantly different (P between different technicians performing the
< 0.0001; Table 2). Patients were followed with test.l,s It also allows the clinician to classify the
is-hCG, estradiol, and progesterone determinations patient into one of two specific groups «14% and
on a weekly basis, pelvic ultrasound starting on the > 14% normal morphology), giving a reliable crite-
seventh week after the last menstrual period, and rion that can be used to counsel the patient and to
clinical evaluation to determine pregnancy status plan the approach in future IVF cycles.
and type of gestation. The pregnancy rate in group Based on the significant differences between
I was lout of 13 patients (7.6%) and in group II was normal morphology and the slightly amorphous
10 out of 32 patients (31.2%). The ongoing preg- forms in groups I and II, we propose that two pat-
nancy rate in group I was lout of 13 patients terns can be observed in the <14% normal mor-
(7.6%) and in group II, 6 out of 32 patients (18.7%), phology group. The P pattern (poor prognosis pat-
with three clinical miscarriages and one ectopic tern) has a mean normal morphology of 1.8% and
pregnancy. The differences between these two mean slightly amorphous forms of 18%, with a
groups in terms of reproductive performance did morphology index < 30% (Table 1). The G pattern
not reach statistical significance because of the (good prognosis pattern) gives the patient a signifi-
small number of patients. cantly better chance to fertilize (P < 0.0001) than
Of the 45 patients studied, 14 had a hamster test the P pattern (Table 2). The mean normal mor-
performed prior to the IVF procedure. All 14 had a phology in the G pattern was 7.7%, the mean
penetration rate below 10%. Four out of the 14 slightly amorphous forms were 34.3%, with a mor-
(28.6%) patients with poor penetration rates did phology index> 30%. Based on these patterns,
not fertilize any oocytes in vitro, but 10 of 14 pa- predictions on chances of fertilization can be done
tients (71.4%) did fertilize in vitro. Five of 10 pa- with much more accuracy in the group with <14%
tients (50%) fertilized >50% of the eggs, and 5 normal morphology.
(50%) fertilized <50% of the eggs. The fertilization rate for all patients with a nor-
mal morphology < 4% and morphology index
< 30% (P pattern) was 7.6%; when the normal mor-
DISCUSSION
phology was >4 % and the morphology index> 30%
Normal morphology evaluated by strict criteria (G pattern), the fertilization rate was 63.9%. Only
is a valuable tool to predict a patient's chance to one pregnancy was established in patients with a P
fertilize and to reach the transfer stage. In a pre- pattern; in patients with a G pattern, the ongoing
vious study performed at the Jones Institute,4 70 of pregnancy rate was 60%, which compares favorably
71 patients with normal morphology > 14% with the ongoing pregnancy rate previously re-
reached the transfer stage, reflecting a high fertil- ported in our overall population.14 This observation
ization rate in this group. If the normal morphology again confirms previous reports 15 ,16 that if fertiliza-
is <14%, the fertilization rate per oocyte is mark- tion occurs, the performance of the embryos, as
edly impaired. l ,4 This study was designed to evalu- well as the transfer and pregnancy rates, are no
ate the sperm morphology in this group and to try different from the general IVF population.
to define morphologic patterns in patients with and The question now arises whether the fertilization
without fertilization. Our results indicate that se- rate and prognosis of patients with normal mor-
vere impairment of fertilization will take place at a phology < 14% can be improved, especially in those
level of <4% normal morphology, based on the with a P pattern, but also in those with a G pattern
strict criteria explained previously (Table 2). Re- who fertilized <50% of the oocytes. Can they per-

Vol. 49, No.1, January 1988 Kruger et al. Abnormal sperm morphology and IVF 115
 haps benefit by simply increasing the concentra- Acknowledgments. We thank the scientists in the In Vitro
tion of sperm per milliliter of the insemination me- Fertilization Laboratory, Ms. Simona Simonetti, Dr. Jake
dium at the time of IVF from 50,000/ml to Mayer, and Ms. Mary Maloney for making the slides daily, and
Ms. Debbie Jones for retrieval of data. We also thank the tech-
500,000/ml? There have been several reports nicians in the Andrology Laboratory, Mrs. Rosita Acosta, Ms.
warning against a significant decrease in fertiliza- Anne Bogaert, and Ms. Mary Hamilton, for their devoted work,
tion rates in vitro in mice and hamsters when the and Mrs. Sharon Durio and Mrs. Myra Waters for secretarial
sperm concentration was increased. 17,18 This de- assistance. Last, but not least, we thank the SA Medical Re-
crease can be due to excessive numbers of antifer- search Council and Tygerberg Hospital who financially assisted
the first author during his stay at the Jones Institute for Repro-
tilization factors19 or proteases20 near the oocyte. ductive Medicine.
Nevertheless, in the beginning of our own program,
insemination was done routinely with 500,000
sperm/mllegg and the fertilization rate was no dif-
ferent. It also was demonstrated in studies with REFERENCES
mice, using suboptimal concentrations of sperm,
that as the sperm density is reduced, fertilization 1. Kruger TF, Menkveld R, Stander FSH, Lombard CJ, Van
der Merwe JP, van Zyl JA, Smith K: Sperm morphologic
rates also are reduced. 19 To answer these questions, features as a prognostic factor in in vitro fertilization. Fertil
a prospective study is being conducted. Steril 46:118, 1986
Another important point in male factor cases is 2. Mahadevan MM, Trounson AO: The influence of seminal
the timing of insemination. In Norfolk the extru- characteristics on the success rate of human in vitro fertil-
sion of the polar body is used as an indicator of ization. Fertil Steril 42:400, 1984
3. Alper MM, Lee GS, Seibel MM, Smith D, Oskowitz SP,
oocyte maturity,12 at which time insemination Rnasil BJ, Taymor ML: The relationship of semen parame-
takes place. ters to fertilization in patients participating in a program of
The correlation of the sperm penetration assay in vitro fertilization. J In Vitro Fert Embryo Transfer 2:217,
(SPA) and IVF was not good in this study, with 1985
28.6% no fertilization and 71.4% fertilization rate 4. Kruger TF, Acosta AA, Simmons KF, Swanson JR, Matta
JF, Veeck LL, Morshedi M, Brugo S: A new method of
per patient with <10% SPA penetration in all 14 evaluating sperm morphology with predictive value for IVF.
cases. In a previous study21 this group indicated Urology. 30:248, 1987
that there is a good correlation between normal 5. Jeulin C, Feneux D, Serres C, Jouannet P, Guillet-Rosso F,
morphology and SPA penetration rate. If normal Belaisch-Allart J, Frydman R, Testart J: Sperm factors re-
morphology was <14%, 85% of cases did not pene- lated to failure of in-vitro fertilization. J Reprod Fertil
76:735, 1986
trate above the 10% level. If normal morphology 6. Acosta AA, Bernardus RE, Jones GS, Garcia J, Rosenwaks
was >14%, the penetration rate above the 10% Z, Simonetti S, Veeck LL, Jones D: The use of pure FSH
level was 86% (P < 0.0001). We conclude from alone or in combination for ovulation stimulation in in vitro
these observations that the SPA is giving us the fertilization. Acta Eur Fertil 16:81, 1985
same information as normal morphology greater or 7. Van Zyl JA,oMenkveld R, Kotze TJ van W, Van Niekerk
WA: The importance of spermiograms that meet the re-
lesser than 14%1,4 in the population studied. A SPA quirements of international standards and the most impor-
< 10% and a normal morphology < 14% are param- tant factors that influence seinen parameters. In Proceed-
eters warning the clinician of potential problems in ings of the 17th International Society of Urology, Vol 20.
IVF due to the male factor. Without identifying the Paris, Diffusion Doin Editeurs, 1976, p 263
different patterns (P and G), valuable predictive 8. Kruger TF, Ackerman SB, Simmons KF, Swanson RJ,
Brugo S, Acosta AA: A quick reliable staining technique for
information will be lost. sperm morphology. Arch Androl. 18:275, 1987
It is worth emphasizing that these criteria are 9. Menkveld R: The influence Of environmental factors on
useful only in IVF with the techniques used in spermatogenesis and sperm parameters. Doctoral thesis,
Norfolk. The significance of these abnormalities in University of Stellenbosch, Parow, South Africa, 1987
clinical practice remains to be demonstrated. 10. Eliasson R: Standards for investigation of human semen.
Andrologie 3:49, 1971
The evaluation of sperm morphology is a con- 11. World Health Organization: Lahoratory Manual for the Ex-
troversial issue. Results in fertilization rates in IVF amination of Human Semen and Semen-Cervical Mucus
units differ. 1,2,4,16 Do we look at the same spectrum Interaction. Singapore, Press Concern, 1980, p 16
of abnormalities, explaining the difference in re- 12. Veeck LL, Maloney M: Insemination and fertilization. In In
sults, or is our classification of abnormally and Vitro Fertilization, Norfolk, Edited by HW Jones Jr, GS
Jones, GD Hodgen, Z Rosenwaks. Baltimore, Williams &
normally shaped sperm in need of revision? It is Wilkins, 1986, p 168
our opinion that the latter is true and thus needs 13. Swanson RJ, Mayer JF, Jones KH, Lanzendorf SE, Mc-
the attention of those involved in the field. Dowell J: Hamster ova/human sperm penetration: correla-

116 Kruger et aI. Abnormal sperm morphology and IVF Fertility and Sterility
 tion with count, motility, and morphology for in vitro fertil- 18. Talbot P, Franclin LE, Fussel EN: The effect of the con-
ization. Arch Androl 12(suppl):69, 1984 centration of golden hamster spermatozoa on the acrosome
14. Jones HW Jr, Acosta AA, Andrews MC, Garcia JE, Jones reaction and egg penetration in vitro. J Reprod Fertil
GS, Mayer J, McDowell JS, Rosenwaks Z, Sandow BA, 36:429, 1974
Veeck LT, Wilkes C: Three years of in vitro fertilization at 19. Stranger JD, Quinn P: Fertilization of cumulus-free, zona
Norfolk. Fertil Steril42:827, 1984 intact mouse ova in vitro at high and low sperm concentra-
15. Acosta AA, Chillik CF, Brugo S, Ackerman S, Swanson RJ, tions. Gamete Res 5:61, 1982
Pleban P, Yuan J, Haque D: In vitro fertilization and the 20. Vacquier VD, Tegner MJ, Epel D: Protease release from sea
male factor. Urology 28:1, 1986 urchin eggs at fertilization alters the vitelline layer and aids
16. Yovich JL, Stanger JD: The limitations of in vitro fertiliza- in preventing polyspermia. Exp Cell Res 80:111,1973
tion from males with severe oligospermia and abnormal 21. Kruger TF, Hamilton M, Simmons KF, Swanson JR,
sperm morphology. J In Vitro Fert Embryo Transfer 1:172, Acosta AA, Matta JF, Oehninger S, Morshedi M: The sig-
1984 nificance of abnormal sperm morphology and other sperm
17. Iwamatsu T, Chang MC: Factors involved in the fertiliza- parameters in the outcome of the hamster ova/human
tion of mouse eggs in vitro. J Reprod Fertil 26:197, 1971 sperm penetration assay (SPA) Int J Androl. In press

Kruger et al. Abnormal sperm morphology and IVF 117

Vol. 49, No.1, January 1988