| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Andrology Lab Corner |
From BioPore Inc, State College, Pennsylvania.
| Correspondence to: Rupert P. Amann, 819 Marble Dr, Ft Collins, CO 80526 (e-mail: ramann{at}lamar.colostate.edu ). |
| Reprint requests: amann{at}biopore.com . |
| Received for publication November 14, 2001; accepted for publication December 21, 2001. |
For more than 50 years, a dogma has evolved for the study of relationships between sperm quality and fertility, with the latter measured somewhat differently for each study. Typically, a graphic display depicts data pairs for sperm quality and fertility, either as individual data points or after smoothing, and is accompanied by the presentation of a correlation coefficient. Sometimes, a tabular summary is used, with values for sperm quality arbitrarily assigned to at least 2 groups and the associated fertility outcomes averaged within groups. In either case, the assignment of "blame," if the anticipated relationship is not demonstrated, and the emphasis in the discussion when data are published (many "disappointing" studies remain unpublished) usually are placed on the measure of sperm quality. The limits in data providing the fertility outcome are minimized or ignored. We address the inherent limits when measuring the fertility outcome.
In this paper, we focus on the measurement of fertility in situations when it is important to maximize the probability of reaching the right conclusion in initial studies, before moving on to large-scale clinical or commercial applications. Among the countless studies reported in the literature, some were designed to be conclusive. Unfortunately, the data presented in most publications have severe limitations when used, as planned, for the validation of a diagnostic assay or the detection of the effects of a handling process or seminal additive on the fertilizing potential of sperm. We made no attempt to review the thousands of publications. Rather, we present opinions on the measurement of fertility of a male, or semen sample, with respect to the above questions 4 and 5.
Regardless of the reason, the measurement of fertility is not a trivial task for any species. It is especially difficult with males of human or endangered species. We intentionally present concepts as generalities and believe that they are applicable in most cases and with most species, including humans. However, with dairy cattle, other options are available after initial studies, because the IUI of more than 400 females per ejaculate is possible.
Fertilizing Potential and Fertility
Regardless of what is being measured, meaningful outcome data must be based on sufficient numbers of experimental units (eg, females or dishes of eggs) and obtained using defined and validated conditions. "Fertilizing potential" is the probability (or capability) of a spermatozoon or population of spermatozoa in a sample (dose) of semen to successfully participate in fertilization, namely formation of a 1-cell embryo or zygote. Fertilizing potential cannot be directly measured, except via coincubation with oocytes for IVF using appropriate conditions (rarely provided; see below).1 However, IVF does not allow or require the expression of the complete repertoire of attributes that must be functional in sperm deposited into a female by copulation or IUI. Further, all too frequently, fertilization occurs, but the resulting 1-cell embryo does not develop sufficiently to prevent a mammalian female from expressing the next menses or estrus or to deliver a live offspring. A value for fertilizing potential can not be extrapolated from one measure to another.
The term "fertility" is used in veterinary medicine and animal agriculture to report the percentage of females clinically pregnant, avian eggs containing a living embryo, or embryos formed, as detected at a defined point after copulation, IUI, or IVF, using a standard measure of clinical pregnancy or embryo formation. Usually, data are averaged across at least 1 cycle for each of the 10 or more females exposed via IUI to sperm from 1 male (ie, a sperm donor) or across 4 or more oocytes from at least 1 female exposed via IVF to sperm from 1 male (minimum numbers; see Amann 1989). In any case, the term applies to populations of animals.
With most mammalian or avian species, the calculation of population probabilities is both possible and desirable. The interpretation of 1 or 2 missed menses is well known. With cattle or pigs, a "30-day nonreturn rate" would be the percentage of females failing to demonstrate estrus by 30 days after the most recent copulation or insemination (during a previous estrus), assuming the systematic observation of animals for signs of estrus. In human medicine or small-scale animal studies, ultrasonic visualization of the uterine contents is practical. With poultry, eggs often are examined for the presence of a viable embryo, via transillumination, after 4-10 days of incubation; fertility is the percentage of eggs with a live embryo.
The outcome of any copulation, IUI, or IVF depends on the interplay of both male and female gametes, as well as on other female factors, plus the variation introduced by still other factors (known or unknown). Hence, when reporting fertility, the observed fertility really is the product of (male fertility) times (female fertility) times (all other source of variation). Unfortunately, this hampers ascribing the observed result to the male or his sperm. Indeed, in large-scale fertility studies (>20 males and >10 000 females) based on the commercial IUI of pigs or Holstein cattle, the variation associated with "other sources" usually is 10-20 times that associated with either males, seminal treatments, or their interaction (Foote and Oltenacu, 1980; Weigel, 2000; Christensen, personal communication). Obviously, there can be a profound effect of the lifestyle of humans or the husbandry management of animals, parity, age, environment, laws of probability, random variation, etc. Many of these variables can be controlled in small research studies with animals, but this is difficult for human clinical trials. The use of a large data set does not eliminate these confounding factors, but their impact can be minimized mathematically.
Even when ignoring the problem of all other sources of variation, an observed fertility of 0.50 (ie, 50%) could result from male fertility of 0.53 and female fertility of 0.95, equal male and female fertilities of 0.71, male fertility of 0.85 and female fertility of 0.59, or some other combination. The substantial impact of the female and her management or environment can be minimized by the use of heterospermic IUI (Beatty, 1960; Saacke et al, 1982; Dziuk, 1998) or IVF (Blazak et al, 1981). The design and interpretation of heterospermic studies are outside the scope of this paper.
In contrast to populations, the term "fecundability" pertains to a specific couple or pair of animals. Fecundability is the probability that conception will occur within that female, during 1 estrual or menstrual cycle, when exposed to sperm from a specific male. Actually, "recognizable fecundability" is measured, and the criterion of outcome must be stipulated. These latter terms are traditional in human medicine. The impact of the female is acknowledged and explicit and, as with fertility, can affect the outcome regardless of sperm quality. If a man participated as a member of many couples—as a donor of cryopreserved human sperm might—the average recognizable fecundability for that male's sperm across numerous females or cycles of IUI could be calculated; this also could be termed fertility.
Utility of Outcome Data From IUI or IVF
Muller (2000) emphasized that it is fallacious to consider the outcome (ie, not pregnant vs pregnant, by some standard measure) after copulation or IUI an appropriate endpoint for establishing the predictive power of a laboratory diagnostic test of sperm quality. Success or failure is not dependent solely on attributes of sperm quality. Success after IUI is evidence that all of the important attributes associated with sperm quality performed "enough" of the needed functions in a "combined effective amount" in a sufficient subpopulation of sperm (see Amann and Hammerstedt, 1993) at the right time. However, failure can be due to the malfunction of one of many sperm attributes and quite possibly not the one evaluated by the diagnostic laboratory assay. Alternatively, failure can be entirely due to female or other factors.
Although the birth of a live young (vs no live young) is not a useful or appropriate endpoint for establishing the predictive power of a laboratory diagnostic test of sperm, it is an important outcome. Thus, the fact that a given IUI or IVF did or did not result in a live birth likely is the most appropriate measure of outcome when evaluating a new seminal additive or sperm processing procedure. In many cases (eg, large studies with animals), however, it is more practical to measure pregnancy rates ca 2 months after IUI or embryo transfer.
Muller (2000) stated that the outcome (eg, cleavage or formation of an embryo appropriate for transfer) during IVF is the best available endpoint for expressing fertility when obtaining data to validate a potential diagnostic assay for sperm quality. While it may be the best available, it is not ideal as usually implemented. As with IUI, the cause of the failure of sperm to fertilize 1 or more eggs, out of 6-20 (for example), during IVF is not necessarily due to any individual attribute of sperm function. Indeed, the cause of the failure could be different for each unfertilized ovum. Many important sperm attributes are bypassed in IVF and might be defective but undetected. These include attributes involved in sperm survival within the female tract and interaction with the oviductal epithelium. Also, with some species (eg, cattle), nonphysiological approaches are used with IVF to replace in vivo mechanisms for capacitation and stimulation of the acrosome reaction. However, female factors not involving the oocytes can be excluded.
Impact of Procedure to Process Sperm
The impact of procedures used to process sperm before measurement of sperm quality in the laboratory or measurement of fertility by IUI or IVF can not be ignored. Spousal sperm used for IUI or IVF usually are processed (eg, by swim-up or gradient centrifugation) to exclude immotile and abnormal sperm. Similarly, most ejaculates of human donor semen for use in IUI or IVF are processed before freezing to provide a thawed preparation of high quality, ready for IUI or IVF. Such processing introduces a potentially confounding factor unless the diagnostic test is based on the evaluation of sperm samples prepared in a similar manner and not simply neat or diluted semen. This would be easy to build into the protocol for a diagnostic assay but could add extra cost.
Some laboratories include enhancers of sperm function (eg, caffeine, pentoxifylline, progesterone, and ionophore) in the medium used to process sperm or to coincubate sperm and ova. Such enhancers rarely are included in the same concentrations during the laboratory evaluation of an aliquot of the sample or even another sample from the same male.
Impact of Number of Sperm Inseminated or Coincubated With Ova
There is an additional limitation to conventional fertility data. The impact of the number of sperm deposited in the female reproductive tract, by copulation or IUI or coincubated with oocytes during IVF, generally is overlooked although well documented with several species (Figures 1 and 2; Pace et al, 1981; Saacke et al, 1994; Saeki et al, 1995; den Daas, 1998; den Daas et al, 1998; Fearon and Wegener, 2000; Hasler, personal communication). Unfortunately, the number of sperm usually used for IUI or coincubation with mature oocytes for IVF is not constant and generally substantially exceeds the number needed for maximum fertility with sperm from most males. This is: 1) logical, given the goal of maximizing the probability of fertilization of each oocyte, but 2) illogical and contraindicated, if the goal is measuring the potential of sperm in a given sample to fertilize oocytes in vivo or in vitro.
|
|
The valid measurement of the outcome for any response requires that conditions be defined for placing the results on the dose-responsive portion of the measurement curve, typically a sigmoidal, exponential, or hyperbolic curve (Figure 3; Fearon and Wegtener, 2000) if both axes are plotted as measured. This is just as true for fertility as for hormone concentration and is obvious in data for chickens (Figure 4, left panel). Clearly, a limiting number of sperm/IUI must be used to detect differences among populations of males (Figure 4, center panel) that are otherwise undetectable. When a limiting number of sperm was used for IUI, a bimodal distribution was evident (Figure 4, right panel); roosters giving less than 70% fertility likely would have been considered subfertile.
|
|
This important concept is illustrated further in Figure 5. The upper panel shows the distribution of fertility for a hypothetical population of males, representing a hypothetical species, with a typical number of sperm used for IUI or IVF (solid line). The population appears to be bimodal, although the larger part of the population has a substantial tail toward the left. Deconvolution of the hypothetical profile shows, however, that: 1) the larger population includes subpopulations of subfertile males (for simplicity, only 2 are shown); and 2) the distribution of normal males is symmetrical and reasonably narrow. Also evident is the fact that the smaller population, on the extreme left, includes sterile males and others with severe subfertility.
|
The center panel of Figure 5 shows only the normal males and their fertility with a typical number of sperm per IUI or IVF; data are identical to those in the upper panel. The detection of differences among males, or of a change in fertility, would require data for many females (or dishes of eggs) for each male or treatment to reduce the uncertainty (ie, 95% CI) of each mean to a low value (eg, plus or minus 5 percentage units; requiring at least 500 females per male or treatment). This is impossible with humans, although it is possible with some species if data are obtained in a commercial setting using thousands of IUIs. When fertility data are used to validate a diagnostic assay, all fertility data would ideally be obtained using aliquots of the same semen evaluated with the laboratory assay, which is difficult even with dairy cattle.
The lower panel in Figure 5 shows (dashed line) the hypothetical fertility distribution for the same normal males when ca 50% of the typical numbers of sperm are used for IUI or IVF. Far more males have relatively low fertility, and importantly, individuals giving similar fertility with a typical number of sperm now are distinguishable (as was the case in Figure 4). Spreading out the difference among individuals reduces the precision needed for each mean value to establish that 1 male, or 1 treatment, differs from another. A 95% confidence interval of plus or minus 10 percentage units might be sufficient. Hence, the number of females (or dishes of eggs) needed per mean is reduced (eg, to 50; see figure 2 in Amann, 1989); such a study is possible with more species.
Finally, the solid line in the lower panel shows the effect of the exposure of sperm from these same males to a profertility molecule or procedure and inseminated at ca 50% of the typical number. The fertility for many, but not all, males is increased, possibly by 10-20 percentage units. The benefit of the treatment can be detected with important species, as fewer females (or dishes of ova) are needed for each male. The average fertility is increased, but it is unlikely that the maximum fertility would be substantially increased.
For species such as turkeys and pigs, semen is conventionally pooled across at least 4 males. This would hamper the detection of treatment differences using a normal number of sperm per IUI. Again, the use of a reduced number of sperm would facilitate initial studies to determine the impact of a new treatment on fertility.
Importantly, the shape of the plot of fertility vs the number of sperm, as in Figure 3, does not depend only on the individual male (see Figure 1). It also depends on the characteristics of the population of females or ova used to generate the curve; management factors such as correct detection of estrus, the skill of the technicians performing IUI, the timing of the copulation or IUI relative to the onset of estrus or a hormonal signal, and the season or weather. Other features can be manipulated, including media used for IVF, social situation, nutrition, and venereal or other disease status. These are the types of factors that contribute to the "other sources of variation" introduced in "Fertilizing Potential and Fertility."
From the perspective of validating a diagnostic assay, the use of an excessive number of sperm when measuring fertility increases the probability that "compensable defects" in sperm will be masked. A compensable defect is one in which low fertility can be overcome, at least in part, by increasing the number of sperm per IUI (Pace et al, 1981; Saacke et al, 1994, 2000; den Daas, 1998). Low fertility caused by an uncompensable defect persists regardless of the number of sperm per IUI or IVF. Hence, with a compensable defect of sperm, the "problem" causing low fertility results from the failure of a sperm attribute (or attributes) being expressed before sperm enter the oocyte. An uncompensable defect involves an attribute (or attributes) being expressed only after a spermatozoon enters an ovum. As discussed by Saacke et al (2000), uncompensable defects often are evidenced by a slow progression through early cleavage divisions and/or an early embryonic death. When a spermatozoon with an uncompensable defect fertilizes an oocyte after IUI, clinical pregnancy may not be detectable. However, the expression of uncompensable defects could be studied by the transfer into recipients of embryos resulting from IVF or intracytoplasmic sperm injection (ICSI) using gametes from a male thought to produce a high proportion of sperm with uncompensable defects.
We are unaware of published data allowing a firm conclusion with regard to the question if human semen includes sperm with compensable defects, other sperm with uncompensable defects, and some sperm with both types of defects. However, occurance of both compensable and uncompensable defects is likely based on accepted clinical practice and data emerging from the use of ICSI with human oocytes followed by embryo transfer. There are abundant data that this is true for other species, as illustrated in Figures 1 and 4. A mechanism for the compensation within individual cells for certain defects is presented in Amann and Hammerstedt (1993), and the benefit of the insemination of more sperm is well known (eg, Figure 1).
If a compensable defect is associated with the attribute measured in the diagnostic assay, fertility could be high under typical conditions of sperm use, although the outcome in the diagnostic assay predicted that it should be low; one might have a high incidence of false-negative outcomes (ie, excess fertilized oocytes). The accumulation of data near or above the asymptote for the change in fertility as a function of sperm number (Figure 3) is especially inappropriate when evaluating the efficacy of a profertility molecule or beneficial procedure affecting spermatozoa, because a beneficial response can not be detected for "lack of range" on the dose-response curve. Ideally, the appropriate number of sperm to adjust fertility to ca 60% of maximum for that male would be determined before conducting the primary study. Realistically, this seldom is possible.
If fertility is the desired outcome measure for the validation of a diagnostic assay, all IVFs must use 1 standard and limiting number of sperm (eg, 20 000 sperm in 100µL). Alternatively, each male's semen could be evaluated for fertility using each of 2 standard numbers of sperm (eg, 15 000 and 30 000 sperm), although this approach is likely impractical. Another approach would be to measure the outcome most biologically relevant to the sperm attribute measured (eg, intracellular calcium concentration [by flow cytometry] for a simple diagnostic purporting to assay capacitation), although clinicians inevitably will ask how the outcome in the diagnostic assay relates to fertility.
Measure Subfertility or Fertility?
As detailed elsewhere (Amann and Hammerstedt, 1993), a given laboratory test evaluates 1 or 2 attributes of individual members of a population of sperm. Assuming several tests are utilized, it can be established what proportion of the population of sperm is defective in 1 or more attributes. Such cells likely are unable to fertilize an oocyte. Recall that many sperm will be defective in more than 1 attribute. Also, many sperm not detected as defective may, indeed, have a problem that prohibits them from fertilizing an oocyte. Hence, one should think in terms of using a laboratory assay to predict if an individual male is likely to be subfertile (including sterile2) or nonsubfertile (Amann, 2000). This distinction is based on biology and not semantics, and fertile vs nonfertile is inappropriate. Importantly, in certain situations, it might be appropriate to introduce a third classification, namely "indeterminate" (Guzick et al, 2001).
Unfortunately, the review of molecular markers by Braundmeier and Miller (2001) was not framed in recognition of the need to predict subfertility rather than fertility (Amann and Hammerstedt, 1993). Also techniques of prediction rather than correlation should be used (McNeil et al, 1975; Zaneveld and Jeyendran, 1992; Amann, 2000). As emphasized by Muller (2000), of the alternative outcomes after copulation or IUI, nonpregnant or pregnant, pregnant is a better estimate of the normalcy of sperm function than nonpregnant is of abnormal sperm function. After IVF, the percentage of ova cleaving vs not cleaving is a logical endpoint. Cleavage probably is the best measure for fertility in IVF, but passage through 2-3 cleavage divisions is not predictive that an embryo will complete embryogenesis and fetal development. Regardless of the measure of fertility, one should use outcome data from a laboratory assay to predict subfertile males, or samples, and fertility data to establish the correctness of those predictions to identify males with a high proportion of nonpregnant or noncleaving outcomes.
Detecting an Improvement in Fertility
Molecules with the potential to enhance the fertilizing ability of sperm are under development. After appropriate preclinical research, clinical studies will be needed to establish if the fertilizing potential of sperm exposed to such molecules indeed is increased, relative to control samples. IVF would be an appropriate first approach for molecules thought to increase binding to the zona pellucida or facilitate timely acrosome reactions. IUI is the only logical approach for molecules thought to stabilize plasma membranes or enhance the interaction of sperm with oviduct epithelium and, ultimately, for any molecule. Profertility molecules would have the greatest impact with spousal or donor IUI. As noted in "Utility of Outcome Data From IUI or IVF," an appropriate measure of success would be the birth of more live young. It is not inevitable that this must result from an increased fertilization rate.
Typically, the number of sperm in an IVF droplet or IUI dose is substantially in excess of the minimum number needed to maximize "success," although it is set to minimize polyspermy. This maximizes fertility and minimizes the male effect thereon. However, this approach substantially reduces the probability of reaching the right conclusion concerning the efficacy of profertility molecules. To reach the right conclusion in early fertility studies, both "normal controls" and "low-number controls" (untreated) should be used, with appropriate amounts of molecule added to low-number preparations (Figure 3). The low number should be no more than 0.5-0.1 times the normal value. The impact of a profertility molecule can then be detected (Figure 5, bottom panel). This approach was effective for IVF with mouse gametes (Magargee et al, 2000). After the efficacy of a molecule is shown, the profertility effect can be confirmed using larger studies and a more typical sperm dose in IVF or IUI.
With cattle or chickens, the effect of the number of sperm/IUI on fertility can be determined for individual males; curves differ among males (Figure 1; Amann and Hammerstedt, 1993; den Daas et al, 1998). With humans, such dose-response curves can not be obtained. With cattle, IUI with 0.5 x 106 sperm rather than 10-20 x 106 sperm into nulliparous females is practical and reduces fertility to at least 0.6 times normal (eg, Amann et al, 1999). For humans, a strategy could be based on available data. An IUI with at least 20 x 106 thawed donor sperm is considered appropriate and results in a range of fecundity rates, showing the impact of a donor. To balance the goals of detecting treatment differences and establishing pregnancies, the low number for humans might be set at 8-10 x 106 spousal sperm. Alternating the use of treated and control low-number insemination doses in 2 successive cycles, followed by an IUI of the normal number in the third cycle, might be appropriate; informed consent would be essential.
Number and Quality of Observations
Many consider cattle the ideal mammal to study relationships among attributes of sperm quality and establishment of pregnancy after IUI; more than 400 inseminations per ejaculate are routine. Both industry-wide data summaries and research studies are based on the IUI of many females with semen from each of a number of males, plus statistical corrections (usually ignoring the number of sperm/IUI), in an attempt to provide the best estimate of fertility. Nevertheless, bulls account for less than 1% of the total variation, and the variation associated with ejaculate within a bull is two-thirds that among all bulls (Foote and Oltenacu, 1980; Weigel, 2000; Christensen, personal communication). For the reasons summarized immediately below, commercial cattle data have limited utility for linkage with the outcome of sperm quality evaluations.
First, for commercial studs, any subfertile bull is detected after the first 200-500 inseminations and is culled by less than 1.8 years of age. Their semen rarely is used by individuals attempting to relate diagnostic assays and fertility, and they are not included in published summaries such as estimated relative conception rate (ERCR) values, which are calculated quarterly (see www.drms.org/sire ) in the United States and are based primarily on data for bulls older than about 5 yr. The ERCR values are deviations from the population mean (ie, percentage units; deviations from a mean of ca 53%). They have a normal distribution (evidence that markedly subfertile animals are not in the population) and usually range from -13 to +11. Values are between -2 and +2 for approximately 77% of bulls. The 95% confidence interval for most bulls is more than 2.2 units, so usually, only bulls differing by ca 3 units can be separated significantly. Only about 2% of bulls have an ERCR value below -4. Further, ERCR values are averages based on all IUIs made with semen from a male over 0.3 to greater than 5 years, rather than IUIs made with, or contemporaneous with, semen evaluated in the laboratory. Clearly, ERCR (or similar) data have very limited utility for linkage with the outcome in laboratory evaluations of sperm aimed at detecting subfertile bulls.
Second, and more important, is the fact that the number of sperm per IUI differs among bulls (see review by Fearon and Wegener, 2000) and is usually above the asymptotic value for the change in pregnancy rate as a function of sperm/IUI (Figure 1) in order to maximize the probability of pregnancy (IUIs to address reason 2 in the introductory material). However, this is contraindicated if the goal is measuring the potential of sperm in a given sample to fertilize an oocyte and provide a viable embryo (reason 4 or 5 in the introductory material). The valid measurement of outcomes after IUI requires an insemination dose that places results on the dose-responsive portion of the curve (Figure 3) for all males for a given population of females and management conditions.
To provide outcome data in which a bull (or treatment imposed on semen) accounts for a substantial proportion of the total variance, one should carefully control extraneous factors, use only nulliparous cattle of 1 breed (minimize the female factor), and use IUI of either 0.8 or 1.5 x 106 sperm. However, this is logistically difficult, as data for at least 20 males and more than 1000 IUIs/bull would be desirable. The best studies of this type (den Daas, 1992; den Daas et al, 1998) provide data for 20 bulls and highlight the difficulty in selecting conditions appropriate for bulls in general.
Impact of These Opinions in the "Big Picture"
Funds are provided for basic or clinical research with the intent that the results will enhance knowledge, but far more importantly, bring benefit to patients, farmers, and/or benefactors. In reproductive biology, private or federal funds most often are used with the implicit if not explicit goal of increasing reproductive efficiency for human satisfaction (or contraception). Reasonably, review panels seek "proof of value" for a potential diagnostic assay, new procedure, or profertility molecule in the form of a "fertility test." Applicants struggle with how to accommodate this anticipated or actual request in an era of decreasing budget size and increasing concern about the ethical aspects of human and animal research. One response is to "piggyback" a test of a potential diagnostic or therapy on an ongoing project or commercial activity or to use an available "test bed" and thereby obtain "data." We personally have authored far too many publications of this sort, as have many others, and with the benefit of hindsight, we would urge extraordinary caution in believing what you read.
It is recognized that research or clinical trials follow an iterative process. A putative diagnostic for subfertility caused by sperm malfunction or malformation, or evaluation of a profertility molecule, would best be tested initially under conditions to answer question 4 or 5 in the introductory material. If evidence for, or a trend toward, utility under these conditions was apparent, then validation under conditions used to answer question 2 or 3 would logically follow.
Too often, it is not recognized that inappropriate data are worse than no data. The experimental design used often precludes valid testing of the (unstated) hypothesis or stated goal(s). Hence, the conclusion presented is drawn without valid experimental support. The impact of such mistakes, in retrospect, can not be estimated easily. At the least, the funds were "wasted." Perhaps more important, erroneous conclusions mislead others in the design of experiments or clinical studies.
Do not attempt to experimentally measure fertility in a situation or species in which this can not be done with rigor. We suggest that if fertility data to validate a diagnostic assay or to demonstrate efficacy of a profertility treatment can not be obtained in a manner incorporating the principles outlined herein, it is far better to have no fertility data. Only this approach will avoid misleading patients or farmers, granting or regulatory agencies, and the public at large. The take-home message is obvious—to validate a diagnostic test, it often will be far better to precisely and directly measure the response forming the basis for the diagnostic test than to imprecisely measure fertility.
Acknowledgments
Suggestions provided by Drs Preben Christensen, John F. Hasler, Richard G. Saacke, and George E. Seidel Jr during the review of drafts of this manuscript are greatly appreciated. However, opinions herein are those of the authors and not the reviewers.
Footnotes
1 Strictly speaking, this is correct. However, with cattle and horses, it is
a common practice to use nonsurgical uterine lavageto to recover the
unfertilized ova or embryos 6-7 days after IUI. Enumeration of accessory sperm
and estimation of early embryonic death are possible (eg,
Saacke et al, 2000). Also see
Wishart and Staines (1999) for
approaches practical with poultry; the egg is laid the day after
fertilization. 
2 The demarcation between subfertile (but having a low probability of an
occasional spermatozoon capable of fertilizing an ovum and producing an
offspring) and sterile is in part a function of the number of attempts at
success. For human males, the demarcation often is hampered by limited
attempts, perhaps 20-25 rather than more than 100. Also see the interaction of
male fertility, female fertility, and other factors as discussed in
"Fertilizing Potential and Fertility." Hence, we prefer to lump
sterile in with subfertile.
References
Amann RP. Can the fertility potential of a seminal sample be
predicted accurately? J Androl.
1989; 10:
89
-98.
Amann RP. Correlation or diagnosis and prediction. Proc 14th Int Congr Anim Reprod. 2000; 2: 69 .
Amann RP, Gill SPS, Hammerstedt RH. Maximizing Genetic Impact of Roosters. State College, Pa: BioPore Inc; 1997 : 15.1-15.9.
Amann RP, Hammerstedt RH. In vitro evaluation of sperm quality: an
opinion. J Androl.
1993; 14:
397
-406.
Amann RP, Seidel GE Jr, Brink ZA. Exposure of thawed bull sperm to
a synthetic peptide before artificial insemination increases fertility.
J Androl. 1999;20:
42
-46.
Beatty RA. Fertility of mixed semen from different rabbits. J Reprod Fertil. 1960; 1: 52 -60.
Blazak WF, Overstreet JW, Katz DF, Hanson FW. A competitive in vitro assay of human sperm fertilizing ability utilizing contrasting fluorescent sperm markers. J Androl. 1981; 3: 165 -171.
Braundmeier AG, Miller DJ. The search is on: finding accurate molecular markers of male fertility. J Dairy Sci. 2001; 84: 1915 -1925.[Abstract]
Chaudhuri D, Wishart GJ, Lake PE, Ravie O. Predicting the fertilizing capacity of avian semen: comparison of a simple colorimetric test with other methods for predicting the fertilizing ability of fowl semen. Br Poult Sci. 1988; 29: 847 -851.[Medline]
den Daas JHG. Laboratory assessment of semen characteristics. Anim Reprod Sci. 1992; 28: 87 -94.
den Daas JHG. Prediction of Bovine Male Fertility [doctoral thesis]. Wageningen: University of Wageningen; 1998.
den Daas JHG, de Jong G, Lansbergen LMTE, van Wagtendonk-de Leeuw AM. The relationship between the number of spermatozoa inseminated and the reproductive efficiency of individual dairy bulls. J Dairy Sci. 1998;81: 1714 -1723.[Abstract]
Dziuk P. Who will the father be? Perspect Biol Med. 1998;41: 439 -445.
Fearon JM, Wegener PT. Relationship between fertility in cattle and the number of inseminated spermatozoa. J Reprod Fertil. 2000;119: 293 -308.
Foote RH, Oltenacu EBA. Increasing fertility in artificial insemination by culling bulls or ejaculates within bulls. In: Proceedings of the 8th Technical Conference on AI. Columbia, Mo: National Association of Animal Breeders; 1980: 6 -12.
Guzick DS, Overstreet JW, Factor-Litvak P, et al. Sperm morphology,
motility, and concentration in fertile and infertile men. N Engl J
Med 2001;345:
1388
-1393.
Magargee SF, Cramer PG, Hammerstedt RH. Increased in vitro binding and fertilizing ability of mouse sperm exposed to a synthetic peptide. Mol Reprod Dev. 2000; 57: 406 -411.[Medline]
McCulloh DH. QA techniques for the IVF program. In: Proceedings of the Eastern Regional Workshop. St. Louis, Mo: American Association of Bioanalysts; 1996: 3.1 -3.21.
McNeil BJ, Keller E, Adelstein SJ. Primer on certain elements of medical decision making. N Engl J Med 1975; 293: 211 -215.[Abstract]
Muller CH. Rationale, interpretation, validation, and uses of sperm function tests. J Androl. 2000; 21: 10 -30.[Medline]
Pace M, Sullivan JJ, Elliott FI, Graham EF, Coulter GH. Effects of
thawing temperature, number of spermatozoa and spermatozoal quality on
fertility of bovine spermatozoa packaged in 0.5-ml French straws. J
Anim Sci. 1981;53:
693
-701.
Robl JM, Dziuk PJ. Influence of the concentration of sperm on the percentage of eggs fertilized for three strains of mice. Gamete Res. 1984;10: 415 -422.
Saacke RG, Dalton JC, Nadir S, Nebel REL, Bame JH. Relationship of seminal traits and insemination time to fertilization rate and embryo quality. Anim Reprod Sci. 2000; 60/61: 663 -677.
Saacke RG, Marshall CE, Vinsion WE, et al. Semen quality and heterospermic insemination in cattle. Proc 9th Int Congr Anim Reprod. 1982;5: 75 -78.
Saacke RG, Nadir S, Nebel RL. Relationship of semen quality to sperm transport, fertilization, and embryo quality in ruminants. Theriogenology. 1994; 41: 45 -50.
Saeki K, Nagao Y, Hoshi M, Nagai M. Effects of heparin, sperm concentration and bull variation on in vitro fertilization of bovine oocytes in a protein-free medium. Theriogenology. 1995; 43: 751 -759.
Sexton TJ. A new poultry semen extender. 1. Effect of extension on the fertility of chicken semen. Poult Sci. 1977; 56: 1443 -1446.[Medline]
Taneja GC, Gowe RS. The effect of dosage of undiluted semen on fertility in two breeds of fowl. Br Poult Sci. 1961; 2: 81 -89.
Weigel KA. Towards national fertility evaluations. In: Proceedings of the 18th Technical Conference on AI. Columbia, Mo: National Association of Animal Breeders; 2000: 50 -54.
Wishart GJ. Quantitation of the fertilizing ability of fresh compared with frozen and thawed fowl spermatozoa. Br Poult Sci. 1985;26: 375 -380.[Medline]
Wishart GJ, Palmer FH. Correlation of the fertilizing ability of semen from individual male fowls with sperm motility and ATP content. Br Poult Sci. 1986; 27: 97 -102.[Medline]
Wishart GJ, Staines HJ. Measuring sperm:egg interaction to access breeding efficiency in chickens and turkeys. Poult Sci. 1999;87: 428 -436.
Zaneveld LJD, Jeyendran RS. Sperm function tests. Infertil Reprod Med Clin North Am. 1992; 3: 353 -371.
This article has been cited by other articles:
 |
|
 |
|
  K. A. Sutton, M. K. Jungnickel, and H. M. Florman A polycystin-1 controls postcopulatory reproductive selection in mice PNAS, June 24, 2008; 105(25): 8661 - 8666. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Christensen, D. Boelling, K. M. Pedersen, I. R. Korsgaard, and J. Jensen Relationship Between Sperm Viability as Determined by Flow Cytometry and Nonreturn Rate of Dairy Bulls J Androl, January 1, 2005; 26(1): 98 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. P. Amann and D. F. Katz Reflections on CASA After 25 Years J Androl, May 1, 2004; 25(3): 317 - 325. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and rate of increase
(ß) vary independently and differ among males (den Daas,
55%) for sperm samples of good or poor quality
when the concentration of post-swimup sperm was increased from 100,00 to
600,00 sperm/ml (
