Professional Documents
Culture Documents
Piis0015028207006085 PDF
Piis0015028207006085 PDF
Despite a few negative reports (1–3), repeated and indepen- allows the staining procedure according to Shorr (14) and
dent studies in assisted reproduction programs showed that a rapid staining technique such as Diff-Quik stain (15).
sperm morphology is significantly correlated with fertiliza-
The Diff-Quik procedure was introduced by Kruger et al.
tion and pregnancy after different techniques of assisted re-
in 1987 (15), and it was found to be comparable with the
production (4–9). Thus, in general there is only little doubt
results of the Papanicolaou staining method. On the basis
that normal sperm morphology, as evaluated by strict crite-
of these results, the 1992 WHO manual (16) introduced the
ria, is a good predictor for assisted reproduction outcome
Diff-Quik method together with a number of other staining
(10). As a result, sperm morphology represents a corner-
techniques as an alternative to the Papanicolaou technique.
stone in modern andrological diagnostics on which the
Since then the Diff-Quik staining technique has been com-
therapeutic concept for the treatment of such patients is de-
pared repeatedly with other staining techniques as reference
veloped.
for the purpose of a more rapid staining procedure or for ob-
For the evaluation of human sperm morphology, different taining optimal contrast for the computer-assisted sperm
staining methods have been described of which the technique morphology evaluation in both human and animal semen
according to Papanicolaou (11) with subsequent evaluation samples (17–22). Another rapid sperm staining method, in-
according to strict criteria (12) has been accepted internation- troduced by Schirren et al. (23), is the stain-coated Testsimp-
ally as standard procedure (13). Apart from the Papanicolaou lets slides, which frequently are used mainly in smaller
stain, the World Health Organization (WHO) also mentions/ centers and practices for assisted reproduction because of
time constraints. In the survey conducted by Ombelet et al.
(24), from 170 respondents worldwide, 33.5% used the Papa-
Received January 19, 2007; revised March 2, 2007; accepted March 5, nicolaou method, 22.9% Diff-Quik, 6.5% Testsimplets,
2007.
and 4.7% Shorr stain as their routine staining procedure. Ad-
Reprint requests: Ralf Henkel, Ph.D., Department of Medical Biosciences,
University of the Western Cape, Bellville, South Africa (FAX: vantages of the Testsimplets slides are that no fixation step is
27-21-9593125; E-mail: rhenkel@uwc.ac.za). necessary and preparation can be done outside a fume cabinet.
0015-0282/08/$34.00 Fertility and Sterility Vol. 89, No. 2, February 2008 449
doi:10.1016/j.fertnstert.2007.03.027 Copyright ª2008 American Society for Reproductive Medicine, Published by Elsevier Inc.
The lack of information in the recent literature about of the two techniques. This procedure identifies methods that
Testsimplets made necessary a thorough comparison with are interchangeable if the differences within mean 1.96
standard methods including a comprehensive statistical analy- SD are not clinically important. The plot is useful in reveal-
sis. Thus, paying tribute to the importance of an accurate ing relationships between the differences and the averages,
evaluation of sperm morphology in a clinical setup, this study looking for systematic biases and identifying possible out-
aimed at comparing the internationally accepted standard liers. In comparison with the Bland and Altman plot, the
method of strict criteria after staining according to Papanico- mountain plot has the advantages of comparing distributions
laou with the Düsseldorf criteria (25) after Shorr staining and more easily and of finding the central 95% of the data, even
the morphology evaluation after the staining procedure with in case the data are not normally distributed. On the other
Testsimplets. hand, the Passing and Bablok regression does not need these
special assumptions regarding distribution and measurement
errors.
MATERIALS AND METHODS
To compare three commonly used staining techniques for the RESULTS
evaluation of human sperm morphology, we used freshly
Figure 1 shows the typical staining patterns for the three dif-
ejaculated semen samples from a total of 94 patients attend-
ferent staining techniques used. The increased background
ing the andrological outpatient clinics at the Departments of
staining on the Testsimplets slides is also clearly visible
Dermatology and Urology, University of Jena, Germany, ac-
(Fig. 1C). The mean sperm concentration, total motility
cording to standard ethical guidelines. After performing
(WHO AþBþC), and progressive motility (WHO AþB)
standard semen analysis according to WHO guidelines (13)
were 79.1 125.5 106/mL, 43.1% 20.7%, and 33.2%
three smears of each ejaculate were made to be stained ac-
19.0%, respectively. A summary of mean, SD, median,
cording to Papanicolaou (12, 13) and Shorr (13, 14) and
and the coefficient of variation (CV) of the parameters for
with a rapid staining technique, Testsimplets (Waldeck,
the staining methods investigated is depicted in Table 1. Gen-
Münster, Germany). In brief, droplets of 20 mL ejaculate
erally, although the results of the Papanicolaou and Shorr
were smeared on precleaned glass slides air dried for 24
stain were closer to each other, the Testsimplets results de-
hours after which Papanicolaou and Shorr staining was per-
viated. Except for the head defects, the CVs for the Testsimp-
formed. In contrast, for the rapid staining technique using the
lets stain were markedly higher.
Testsimplets, the 20 mL droplet of ejaculate was spread over
the Testsimplets slide by putting on a coverslip, which is in- All morphological parameters (percent normal morphol-
cluded in the kit, and left standing for 30 to 60 minutes at ogy, percent head, midpiece, and flagellar abnormalities)
room temperature. correlated statistically significantly positively among the
different techniques (P<.004) (r ¼ 0.318 to r ¼ 0.797), yet
The evaluations were performed according to strict criteria
with lower correlation coefficients for correlations with the
(12) for Papanicolaou- and Testsimplets-stained smears
Testsimplets (r ¼ 0.318 to r ¼ 0.704). However, comparing
and according to Düsseldorf criteria (25) for Shorr-stained
the mean for percentage normal sperm morphology deter-
smears. Slides were evaluated for the percentages of morpho-
mined by means of the Papanicolaou (4.78% 2.54%) and
logical normal spermatozoa, head defects, midpiece defects,
the Shorr stain (4.75% 2.64%), the Testsimplets proce-
and flagellum defects with a bright-field microscope (Zeiss,
dure resulted in a significantly lower result (3.89%
Jena, Germany) under 1,000 magnification and included
2.53%) (P¼.0044 and P¼.0055, respectively). Except for
evaluation of 200 spermatozoa in the four corners and center
the comparison of the midpiece defects between Shorr and
of each slide.
Testsimplets methods (P¼.1565), all other comparisons
Statistical analysis was performed with use of MedCalc showed statistically significant (P<.01) deviations between
software (version 9.2.0.1; MedCalc, Mariakerke, Belgium). the Testsimplets method results and the results of the Papa-
After testing for normal distribution of the results by means nicolaou and Shorr methods. On the other hand, Papanico-
of the c2 test, the statistical comparison of the three methods laou- and Shorr-staining results did not differ in any respect
was performed by means of the Wilcoxon test, as well as the for any of the analyzed parameters (Table 1).
methods according to Bland and Altman (26) and Passing and
The deviations from the standard techniques by using the
Bablok (27) who described a linear regression method with
Testsimplets method were even more evident after perform-
no special assumptions regarding the distribution of the sam-
ing Bland-Altman plots (Fig. 2A–2C ¼ normal sperm
ples and the measurement errors. In addition, mountain plots,
morphology; Fig. 2D–2F ¼ flagellar defects). For the
which are created by computing a percentile for each ranked
comparison of the Papanicolaou with the Shorr stain for nor-
difference between a method and a reference method, were
mal sperm morphology (Fig. 2A) the distribution of the data
performed (28).
points is more or less evenly spread along the abscissa. How-
The three latter statistical methods are methods that com- ever, a much more triangular shape of the distribution pattern
pare two or more measurement techniques. The Bland and of the data is obvious for the comparisons with the Testsimp-
Altman plot is a simple graphic method by which the differ- lets (Fig. 2B, 2C). This triangular shape of the distribution
ences between two techniques are plotted against the average pattern is indicative that the variation depends on the
450 Henkel et al. Comparison of staining methods for sperm morphology Vol. 89, No. 2, February 2008
Fertility and Sterility
FIGURE 1
Henkel. Comparison of staining methods for sperm morphology. Fertil Steril 2008.
TABLE 1
Summary statistics of morphology evaluation of 94 samples stained according to Papanicolaou and Shorr and with Testsimplets slides.
Papanicolaou stain Shorr stain Testsimplets slides Differences between
Testsimplets and
Papanicolaou/Shorr
Parameter Mean SD Median CV Mean SD Median CV Mean SD Median CV staining (P values)
Normal morphology 4.78 2.54 4.20 53.3 4.75 2.64 4.45 55.5 3.89 2.53 3.62 65.0 .0044/.0055
Head defects 85.64 10.40 88.00 12.2 86.52 9.00 88.75 10.4 88.75 7.28 90.05 8.2 .0047/.0120
Midpiece defects 38.20 12.38 37.00 32.4 36.50 11.86 35.50 32.5 33.78 13.43 32.00 39.8 .0074/.1565
Flagellum defects 22.67 14.89 19.00 65.7 21.78 13.38 18.60 61.4 10.49 11.80 6.10 112.5 < .0001/< .0001
Note: Although significant differences between the rapid staining technique (Testsimplets) and the standard techniques are found for all parameters, there were no
differences between the Papanicolaou and the Shorr stains.
Henkel. Comparison of staining methods for sperm morphology. Fertil Steril 2008.
451
method revealed good correspondence between Papanico- (P<.01). Thus, only the Papanicolaou and the Shorr staining
laou and Shorr stains (Fig. 3). This is obvious by the narrow procedures are unbiased.
symmetrical distribution of the plot around zero. However,
for the Testsimplets stain a marked deviation from zero, DISCUSSION
particularly for flagellar defects, is obvious (Fig. 3B),
The results of this study show an extensive agreement
indicating that the two methods are not unbiased. Moreover,
between the morphology evaluation on Papanicolaou- and
the long tail to the right-hand side in the Testsimplets
Shorr-stained smears evaluated according to strict criteria
plot is indicative of large differences between the two
and the Düsseldorf classification, respectively, with relatively
methods.
high correlation coefficients. The good correlation obtained
Calculation of a Passing-Bablok regression revealed a between Shorr-stained smears and Papanicolaou-stained
significant deviation from linearity for flagellar defects smears could be expected, even if evaluated according to
when comparing the Papanicolaou stain with Testsimplets the Düsseldorf and strict criteria methods, respectively.
FIGURE 2
Bland-Altman plots for the comparison of the three techniques regarding normal sperm morphology (A–C)
and flagellar defects (D–F). The deviation for the Testsimplets method is obvious, especially for flagellar defects.
In addition, B, C, E, and F for the comparison with the Testsimplets show that the variation strongly depends
on the magnitude of the measurement. N ¼ 94.
8 10
6
A Papanicolaou % - Testsimplets % 8
B
Papanicolaou % - Shorr %
+1.96 SD +1.96 SD
4 6
4.1 5.9
2 4
Mean
0 2 Mean
0.0
-2 0 0.8
-1.96 SD
-4 -2
-4.0
-6 -4 -1.96 SD
-4.2
-8 -6
-10 -8
0 2 4 6 8 10 12 0 2 4 6 8 10 12
AVERAGE of Papanicolaou and Shorr (%) AVERAGE of Papanicolaou and Testsimplets (%)
10 25
C D
8 20
Papanicolaou % - Shorr %
Shorr % - Testsimplets %
+1.96 SD
6 +1.96 15
14.3
5.6 10
4
5
2 Mean Mean
0
0 0.7 -0.1
-5
-2
-10
-1.96 -1.96 SD
-4 -15
-4.1 -14.6
-6 -20
-8 -25
0 2 4 6 8 10 12 14 0 10 20 30 40 50 60 70
AVERAGE of Shorr and Testsimplets AVERAGE of Papanicolaou and Shorr (%)
70
Papanicolaou % - Testsimplets %
40
60
E F
+1.96 SD
Shorr % - Testsimplets %
30
50 29.2
20
40
+1.96 SD Mean
30 10 11.2
31.8
20 0
Mean
-1.96 SD
10 12.1 -10 -6.9
0
-1.96 SD
-20
-10 -7.7
-20 -30
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
AVERAGE of Papanicolaou and Testsimplets (%) AVERAGE of Shorr and Testsimplets (%)
Henkel. Comparison of staining methods for sperm morphology. Fertil Steril 2008.
452 Henkel et al. Comparison of staining methods for sperm morphology Vol. 89, No. 2, February 2008
was too deep to allow sharp definition of the outlines of epithe-
FIGURE 3
lium cells, especially if the smears were thick (11). This is also
Mountain plots for the comparison of the Shorr and true for the staining of semen smears where a more intense
Testsimplets methods with the Papanicolaou stain. staining is obtained with the Shorr stain and the outlines of
(A) Normal sperm morphology; (B) flagellar defects. spermatozoa may appear a little fuzzy if the smears are over-
Whereas the Shorr and Papanicolaou methods stained, but otherwise it gives the same differentiations of the
correspond quite well with each other, the deviation different spermatozoa regions as Papanicolaou staining.
from zero for the Testsimplets stain results That the Düsseldorf classification was used should also
compared with the Papanicolaou stain results as pose no problem, because for the identification of morpholog-
reference shows that the two methods are not ically normal spermatozoa this method is basically the same
unbiased. Moreover, the long tail to the right-hand as that for strict criteria. Hofmann et al. (25) describe the ini-
side of the Testsimplets plot in B is indicative of tial Düsseldorf definition of morphological normal spermato-
large differences between the two methods, zoa as more accurate than the WHO definition and close to the
underlining the underscoring of flagellar defects with succeeding strict Tygerberg criteria, but it included sperm
the Testsimplets method. heads with slightly reduced acrosomal areas as still normal.
50
Clinical observations led to the improved Düsseldorf classifi-
cation where spermatozoa with normal heads but pointed
A postacrosomal forms and spermatozoa with acrosome areas
40 Shorr <40% are considered as being abnormal, bringing it even
closer to strict criteria. Our results are also in line with those
Testsimplets
found by Meschede et al. (29) who observed that the mean
30 percentage of morphologically normal spermatozoa was
Percentile
454 Henkel et al. Comparison of staining methods for sperm morphology Vol. 89, No. 2, February 2008
21. Soler C, de Monserrat JJ, Gutierrez R, Nunez J, Nunez M, Sancho M, 31. Root Kustritz MV, Olson PN, Johnston SD, Root TK. The effects of
et al. Use of the Sperm-Class Analyser for objective assessment of stains and investigators on assessment of morphology of canine sperma-
human sperm morphology. Int J Androl 2003;26:262–70. tozoa. J Am Anim Hosp Assoc 1998;34:348–52.
22. Soler C, Gadea B, Soler AJ, Fernandez-Santos MR, Esteso MC, Nunez J, 32. Coetzee K, Kruger TF, Vandendael A, de Villiers A, Lombard CJ.
et al. Comparison of three different staining methods for the assessment Comparison of two staining and evaluation methods used for computer-
of epididymal red deer sperm morphometry by computerized analysis ized human sperm morphology evaluations. Andrologia 1997;29:
with ISAS. Theriogenology 2005;64:1236–43. 133–5.
23. Schirren C, Eckhardt U, Jachczik R, Carstensen CA. Morphological dif- 33. Ragni G, Marzioli S, Levenberg A, Guercilena S. Comparison of the var-
ferentiation of human spermatozoa with Testsimplets slides. Andrologia ious techniques of identifying human spermatozoa morphology. Acta
1977;9:191–2. Eur Fertil 1984;15:437–40.
24. Ombelet W, Pollet H, Bosmans E, Vereecken A. Results of a question- 34. Calamera JC, Vilar O. Comparative study of sperm morphology with
naire on sperm morphology assessment. Hum Reprod 1997;12:1015–20. three different staining procedures. Andrologia 1979;11:255–8.
25. Hofmann N, Hilscher B, Morchen B, Schuppe HC, Bielfeld P. Compar- 35. Hellenkemper B, Cooper TG. Spermienmorphologie: Untersuchung mit
ative studies on various modes of classification of morphology of sperm Hilfe von Testsimplets. MTA Dialog 2005;4:234–6.
heads and results in in vitro fertilization—a preliminary report. Androlo- 36. Schoenfeld C, Amelar RD, Dubin L, Amelar S. A new staining technique
gia 1995;27:19–23. for the rapid determination of the morphologic characteristics of sperm.
26. Bland JM, Altman DG. Statistical methods for assessing agreement Fertil Steril 1981;36:408–10.
between two methods of clinical measurement. Lancet 1986;1:307–10. 37. Schütte B. Human spermatozoa stained with toluidine blue–pyronine.
27. Passing H, Bablok W. A new biometrical procedure for testing the equal- A rapid method for differentiation. Andrologia 1986;18:567–9.
ity of measurements from two different analytical methods. Application 38. Franken DR. To the editors [letter]. Andrologia 1978;10:413.
of linear regression procedures for method comparison studies in clinical 39. Roll P, Rief M, Dirnhofer R. Eine einfache Methode zum Mikroskopi-
chemistry, part I. J Clin Chem Clin Biochem 1983;21:709–20. schen Nachweis menschlicher Samenspuren mittels Testssimplets. Ärztl
28. Krouwer JS, Monti KL. A simple, graphical method to evaluate labora- Lab 1986;32:125–6.
tory assays. Eur J Clin Chem Clin Biochem 1995;33:525–7. 40. Oettle EE. Using a new acrosome stain to evaluate sperm morphology.
29. Meschede D, Keck C, Zander M, Cooper TG, Yeung CH, Nieschlag E. Vet Med 1986;81:263–6.
Influence of three different preparation techniques on the results of 41. Ombelet W, Menkveld R, Kruger TF, Steeno O. Sperm morphology as-
human sperm morphology analysis. Int J Androl 1993;16:362–9. sessment: historical review in relation to fertility. Hum Reprod Update
30. Hidalgo M, Rodriguez I, Dorado J. Influence of staining and sampling 1995;1:543–57.
procedures on goat sperm morphometry using the Sperm Class Analyzer. 42. Keel BA. How reliable are results from the semen analysis? Fertil Steril
Theriogenology 2006;66:996–1003. 2004;82:41–4.