Evaluation of sperm head shape at high magnification revealed correlation of sperm DNA fragmentation with aberrant head ellipticity and angularity

Abnormal head shape of fresh motile spermatozoa evaluated by elliptic Fourier analysis reflected sperm DNA fragmentation, providing a rationale for selecting spermatozoa by microscopic shape assessment in the IMSI procedure.

Hiroki Utsuno, Ph.D., Kenji Oka, M.D., Ayako Yamamoto, B.A., Tanri Shiozawa, M.D., Ph.D.

Volume 99, Issue 6, Pages 1573-1580.e1, May 2013


To test for association between DNA fragmentation and head shape at high magnification in fresh motile spermatozoa.

Observational study.

Academic tertiary care center.

A total of 60 men in our assisted reproductive program.

Quantifying sperm head shape using elliptic Fourier analysis, and detecting DNA fragmentation by use of a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay.

Main Outcome Measures:
Correlation between percentage of spermatozoa with abnormal head shape and percentage of DNA fragmentation.

Elliptic Fourier analysis decomposed sperm head shape into four quantitative parameters: ellipticity, antero-posterior (AP) symmetry, lateral symmetry and angularity. DNA fragmentation was significantly correlated with abnormal angularity, and moderately with abnormal ellipticity, but not with abnormal AP symmetry and lateral symmetry. Forward stepwise multiple logistic regression analysis revealed significantly higher percentage of DNA fragmentation in spermatozoa with abnormal ellipticity and abnormal angularity than in spermatozoa with normal-shaped head (6.1% and 5.4% vs. 2.8%). Spermatozoa with large nuclear vacuoles also correlated with sperm DNA fragmentation, and had si 38 gnificantly higher percentage of DNA fragmentation (4.7%).

Among morphologic features of sperm head, abnormal ellipticity, angularity and large nuclear vacuoles would associate with DNA fragmentation.

  • Gabor Huszar, MD

    In 60 husbands of couples undergoing IVF treatment, the potential association between human sperm head shape and DNA fragmentation was studied with elliptic Fourier analysis and TUNEL assay, respectively. In certain comparisons correlations were found. The authors suggest that these results will be useful in sperm selection for IMSI.

    In this good effort the authors should consider a few additional points: For instance a) The samples selected for the study are in the 200 million sperm/ml. semen range and high motility, not usually IMSI material; b) Experiment 2 which is presented as “swim-up sperm” is not an accurate definition. It is swim up preparation from an Isolate density gradient pellet. Isolate gradients filter out spermatozoa with developmental arrest and cytoplasmic retention, a sperm fraction that is known to feature sperm with abnormal morphology and DNA fragmentation. Thus, changes in sperm parameters are expected not because swim up step , but due to the density gradient centrifugation; 3) TUNEL is a limited approach and does not accurately reflect overall DNA fragmentation. It is likely that the low proportion of sperm with DNA fragmentation in this report is related to TUNEL assessment. Incidentally, the authors
    call sperm heads with large nuclear vacuoles if they are > 4% head area, which is not a value but a cutoff limit. Indeed, Others classify vacuoles “large” in the >13-15% range.

  • jim hotaling

    This paper represents an important step toward extracting some sort of clinically relevant meaning from sperm morphology in a mathematically robust way as opposed to trying to standardize human evaluation of sperm which is problematic. While researchers at Stanford sequenced the genome of a single human sperm earlier this year, they did not manage to do so without destroying it. Until we can do this and extract clinically meaningful information from the monstrous amount of information, we are beholden to other metrics of assessing the quality of individual sperm in vitro for ICSI instead of bulk semen parameters found in vivo. My main question for the authors is whether they controlled for multiple comparisons given the several metrics they used to assess sperm head shape. Further, I am curious what the units are for the OR of DNA fragmentation in table 2, if they are a 1% change in DNA fragmentation, do they believe this is ciinically significant as opposed to statistically?

    • Hiroki Utsuno

      Thank you for the comments and questions.
      We did not controlled for significance levels for the multiple comparisons in table 1. If we applied some collection methods, the p-value would be around the collected alpha-level. However, the significant variables were also selected by stepwise forward selection in the multiple logistic regression analysis in table 2.
      The OR of DNA fragmentation indicated the ratio of the odds of DNA fragmentation in the abnormal and normal shape. The OR thus does not have unit.

      • jim hotaling

        Thank you for reply. I understand that the OR does not have a unit but still find the table confusing. Below is the text from the table. Are you stating that above the cut-off for each value there is an Odds ratio of XX for chances of having increased DNA fragmentation compared to referrant group of 2.8% with normal head shape, thus your odds ratio of 2.19 for abnormal ellipticity would indicate that there is a 2.2 fold increased chance of having DNA fragmentation greater than 2.8% which is statistically significant but if difference was 3.1% would it be clinically significant? What I am driving at is whether the authors think that the changes that their head shape analysis is detecting are clinically significant? Regardless, this work is an important step forward.

        OR . odds ratio for DNA fragmentation at the abnormal threshold of the variable; P . probability of the 95% CI included zero.

  • Ettore Caroppo

    I totally agree with Dr Schlegel when he writes ” They evaluate morphology and DNA fragmentation but do not indicate how often the morphologic assessment identified sperm with DNA fragmentation (or missed DNA fragmentation in morphologically normal sperm). These data should allow a “positive predictive value” for any morphologic assessment. In addition, one has to wonder whether “abnormal sperm,” as identified with these morphologic parameters, have any reproductive compromise since no functional outcomes are reported in the study”.

  • Fernando Bronet, PhD

    In the last years have been published a few articles about IMSI, the quantification of sperm head shape in four quantitative parameters is an interesting approach. The correlation between head shape and DNA fragmentation index has been studied before; in this case the results show that there is a clear relationship. However, DNA fragmentation index is very low in all groups and it would be interesting to study the behaviour in samples that show a high DNA fragmentation index.

    • Hiroki Utsuno

      I agree with the importance for examining samples that show a high DNA fragmentation index. Thank you for your comments and we are planning to present some relationships between head shape and sperm quality in those samples.

  • This study gives additional support to the use of the most “normal” appearing sperm possible for IVF/ICSI. I understand that it is important to discern if sperm head morphology can be used to predict or correlate to degree of DNA damage. But, is there any IVF lab that would be willing to do ICSI on the basis of morphology alone without running concurrent DNA testing? Probably not.

    • Edmund

      Most ivf labs do only use morphology for selection and not dna!

      • Good to know!

        • in practical use what happens is you need to pick a single living sperm which is really only done on morphology while the DNA testing is not done clinically on single living sperm cells. The best intermediate we have now are the binding methods which identify a subset of sperm on HA binding then allow that one to be picked up

          • Is there research in pre sperm selection DNA testing that would result in continued sperm viability for ICSI (i.e. for men with known genetic defects like Klinefelters, extra Y chromosomes, etc.)? Or is it easier for pre implantation genetic testing?

            • direct single cell sperm DNA analysis in real time to allow testing selection and ICSi does not yet exist

              • Craig Niederberger

                Hear hear. The problem is that we don’t have a handle on what morphologic assessment means and how to apply a population based assessment like DNA integrity to decision making in ART. The only reasonable expectation to draw at this point in time is that selecting gametes with good reproductive potential for use in IVF/ICSI is a nearly random process, which needless to say goes against the grain of well intentioned medical intervention.

Translate »