Do we really know what ‘fertile’ and ‘infertile’ sperm look like?
Sperm morphology has always been a most contentious area of male fertility diagnostics. Not least because the laboratory assay is highly subjective and difficult to reproduce but also because of the lack of convincing clinical data over the years. More to the point, as an industry we think we know what a ‘normal’ sperm looks like (according to previously published criteria) but by current laboratory methods will never know whether morphologically normal = fertile. Likewise, we don’t know whether those sperm defined as ‘abnormal’ by those same criteria are in fact ‘infertile’. This is because assays used to assess morphology are associated with a high level of measurement uncertainty.
Why are current assays associated with a high level of uncertainty?
1. Assessments are not in ‘real time’ - examination of fixed-stained sperm on a slide allows improved visualisation of sperm size, shape and structure but of course does not allow this in association with function (essentially motility). This leads to a number of unanswerable questions which will exacerbate the level of associated uncertainty, in that: sperm identified as ‘normal’ by shape and size alone
(i) may not have been motile (before drying/fixing/staining)
(ii) may not have been viable “ “ “
Therefore in statistical terms, we might be attempting to relate pregnancy to the % of sperm with an ideal shape/size but with no idea of their viability or functional status.
In an ideal world, we would analyse the morphology of sperm that retain some functional capacity e.g. viability or motility as in the example below. In this video the sperm on the left may be the ONLY sperm considered to be morphologically normal but can we be certain that this is a FERTILE sperm? and can we rule out the others as infertile with any degree of CERTAINTY? We may assume the sperm in the middle (top) to be functionally defective since it has a broken neck which MAY affect its ability to reach the site of fertilisation and bind to the oocyte. The remaining two (right and bottom) would be considered 'abnormal' by technicians using strict criteria but the question remains as to whether there sufficient (or any) published evidence which would firmly demonstrate that neither possess fertilising capacity? If the answer is NO, which is more than highly likely, then there is no other conclusion to make other than our current methodology is flawed and clinically weak.
An analogy might be in deciding which is the optimum aerodynamic shape/ size for a fast car, deciding on the features and dimensions and then only finding later that the car had no engine i.e. the optimum shape/size becomes irrelevant because the car cannot function.
2. Light microscopy is limited in its ability to resolve ultrastructural defects Some defects at the ultra-structural level e.g. mitochondrial defects, microtubular/central sheath defects, loss of dynein arms (prevents flagellar movement) which will have a direct influence on sperm function cannot be seen with conventional microscopes and need electron microscopy to be detected. This merely adds to the uncertainty created by the issue described above.
3. Current ‘strict’ philosophy assumes that ‘normal forms’ are fertile and ALL sperm which are outside of those strict criteria (including borderline forms) are infertile When in fact neither statement is correct because of the issues described above as well as being significantly open to individual interpretation. This is perhaps the largest potential for uncertainty, particularly because there is no clinical data to support the assumption that only those sperm deemed as being of ‘ideal’ shape (according to strict criteria) are the only sperm that may swim to, and fertilise an egg. Moreover a ‘borderline’ or 'abnormal' form could be created purely as an artefact of slide preparation, fixation or staining.
Take the example below:
A. Globozoospermia
Picture A shows sperm with globozoospermia (round headed defect, no acrosome). Its chance of fertilisation (without assistance) is zero. This will be scored as a defect with good reason and the level of uncertainty near to zero.
A Sperm with globozoospermia
B Amorphous sperm
Picture B shows an amorphous sperm with several nuclei - most likely this will also be scored as a defect with zero chance of fertilisation. Again this is very likely to be a correct decision with a low level of uncertainty
B. Amorphous multinucleate sperm
If a patient’s sample contains a high percentage of those in 1 and 2, we could with a high degree of certainty predict lowered fertility or even sterility.
Picture C however shows a thickened midpiece and head shape that could be described as rounded although still oval. Its chance of being able to fertilize is largely unknown – yet because the assay is categorical and there is no room for ‘uncategorised’ (or borderline), this will also be labelled as ‘abnormal’ and ultimately therefore scored with the same level of risk as those seen in A and B. This is incorrect because sperm A and B possess defects which are categorically associated with poor sperm function. If a patient’s specimen contains a large number of the forms seen in C, the chance of fertilisation would remain largely unknown and the level of doubt (uncertainty) over the entire sample/diagnosis would remain considerable.
C. Thickened midpiece with off-oval head
What is the 'clinical value' of sperm morphology assessment and how is it determined?
The characteristics for the definition of normal and abnormal sperm are related statistically to 'outcome', often by means of complex regression analyses. 'Outcome' may be defined as a natural conception (in which case time factors must be included) or an assisted conception such as artificial insemination or IVF. Those morphological definitions are then analysed for their statistical influence on whether there is a pregnancy (vs no pregnancy) or perhaps a live birth (vs no live birth). If then for example % normal forms is identified as having statistically significant influence on outcome then it could be concluded that the variable also has clinical value. However, logically for any of these statistics to be valid, the normal or abnormal sperm also has to be viable.
Is a TZI or MAI type approach more useful or more valid?
There is some evidence to support that assessing the overall level of defects found in any sperm could be more predictive of fertility. However to assume that this is a significant improvement and associated with a lower level of uncertainty would be a flawed argument since the fundamental problem remains in that we cannot be certain of recognising ‘fertile’ sperm and we know that ALL DEFECTS ARE NOT EQUAL in terms of determining risk of failed sperm transport, binding, acrosome reaction, fertilisation etc
For example – according to current criteria and using the examples above
Classification Uncertainty
Sperm A = Globozoospermia = abnormal LOW - True result sperm is infertile
Sperm B = Multinucleate = abnormal LOW - True result sperm is infertile
Sperm C = thickened midpiece = abnormal HIGH - FALSE result fertility status is unknown
The same might be said for any number of defects categorised as abnormal but where their functional status remains unknown. Therefore, the possibility remains that sperm with for example: off-oval head, pyriform head, elongated head, coiled tails retain fertilising capacity, yet are labelled as abnormal. We also know that some of these defects can simply exist as artefacts of slide preparation (smearing), fixation or staining.
What is the solution?
It would be nice if there was a simple solution but clearly there isn't as the industry has wrestled with the issue for more than 40 years. What is clear is that sperm morphology is important and cannot be ignored, since failing to detect an abnormality which is definitely associated with infertility or sterility as those examples above (A and B) would be deemed negligent. However, due to the reasons given above, performing the assay according to currently recommended methods is of little clinical value and some of the largest most comprehensive studies demonstrate this in relation to ART (see Hotaling et al, 2011 Fertil Steril and Van den Hoven et al Fertil Steril 205). One thing for sure is that if we don’t continue to ask questions and continue to accept the status quo, the situation (clinical value of morphology analysis) will never improve and in some cases diagnostic testing will continue to be misleading.
Some of the very early functional studies (30 years or more ago) examined the morphology of sperm that was found in the female reproductive tract or bound to the oocyte. These approaches were of course more technically challenging and resource hungry but they did demonstrate a clear advantage in that the functional sperm population was separated from the non-functional (clinically redundant) portion before morphology assessment. Perhaps a re-visit of some older science and a move away from the examination of the heterogeneous sperm population within semen is the only way of providing a morphology assessment with reduced uncertainty.