The present sDFI results show variability between the donor stallions in terms of their post-thaw sperm DNA integrity. This is not surprising given the individual variability commonly seen in stallion sperm [2, 3, 32]. Horses have undergone selection primarily with pedigree in mind or for their ability to provide a particular type of service. In some cases, however, this has led to reduce fertility [9, 15, 20, 23, 33]. In the present work, the range of post-thaw sDFI values immediately after thawing (6.55-22.46%) recorded is similar to that reported by López-Fernández et al.  (4.2-26.3%).
Individual variability was observed in the DNA fragmentation dynamics recorded, as reported earlier by López-Fernández et al. . In the present study, the sDFI values (i.e., DNA fragmentation levels) increased significantly with incubation time, as reported by other authors [6, 9, 34]. The largest increase occurred between 6 h and 24 h of incubation, not between 1 h and 6 h as described by López-Fernández et al. . This may be significant since there is increasing evidence that the integrity of the sperm chromatin at the actual time of fertilization influences embryo survival ; poor sperm DNA integrity may account for some of the infertility commonly thought to lie with mares.
Hoogewijs et al.  obtained an average sDFI immediately after thawing of around 12% in colloidally-selected stallion sperm samples cryopreserved with Botu-Crio®, while Cortés-Gutiérrez et al.  reported values of 18-22% in non-colloidally-selected donkey sperm samples frozen with Gent®. The present results are consistent with the findings of these authors (13.24% for B-1, 11.86% for B-2 and 17.1% for G-C), although different DNA integrity evaluation techniques were used (SCSA, N-Comet Assay and the sperm chromatin dispersion test respectively).
No differences were seen in the sDFI between any of the nine treatments immediately after thawing. However, the use of InraFreeze® was associated with a trend towards the improvement of the sDFI values, especially at 24 h of incubation; indeed significant differences (p<0.05) were seen between I-2 and B-C, I-2 and G-C, and I-1 and G-C.
Irrespective of the centrifugation protocol followed, InraFreeze® significantly reduced the sDFI values at 6 h compared to Botu-Crio®, and at 24 h of incubation compared to both Botu-Crio® and Gent®. Some authors report that egg yolk-based extenders better protect chromatin structure than skimmed milk-based extenders, at least in bovine sperm . Carvalho et al. , in contrast, report the extender employed to have no influence on DNA integrity. However, it should be noted that, in these two studies, the sDFI was analysed immediately after thawing; no post-thaw incubation study was undertaken - and DNA integrity would be most strongly affected during such an incubation period. The precise mechanism by which egg yolk aids in the protection of spermatozoa during the freeze-thaw process is unknown , but low density lipoproteins present in egg yolk plasma are widely presumed to be the cryoprotective agent [38, 39]. Further, there is increasing evidence that cryoprotective antagonists may exist in other egg yolk fractions [38, 39]. The above may explain the better sDFI values obtained with InraFreeze®, which contains only egg yolk plasma rather than whole egg yolk as in Botu-Crio® and Gent®. Replacing egg yolk with sterilized egg yolk plasma might remove a potential source of cryoprotective antagonists .
In the present work, single layer colloidal centrifugation prior to cryopreservation (Protocols 1 and 2) led to significantly lower (p<0.05) sDFI values than the SCC protocol immediately after thawing. These data are consistent with those obtained by other authors for human sperm [40, 41], fresh equine semen [19–21, 25], cooled equine semen  and cryopreserved equine semen subjected to colloidal centrifugation before freezing [28, 29] or after thawing [17, 29]. However, these authors performed no DNA fragmentation dynamics analysis.
The present results showed the I-2 and G-2 treatments to return significantly lower sDFI values than the B-C treatment at 4 h of incubation. At 24 h of incubation, Protocol 2 was associated with the best sDFI values for all three extenders. Crespo et al.  found no differences in sDFI values at time 0 of incubation between colloidal, standard or non- centrifuged equine fresh sperm samples, but as we found, after incubation at 37°C samples subjected to colloidal centrifugation exposed lower sDFI values than non-selected sperm both for fresh and cooled equine semen, what can be related to higher longevity in the selected sperm. Macías García et al.  have recently proved that colloidal centrifugation selects a spermatozoa subpopulation that clearly responds differently to osmotic shock, which could better withstand cooling procedures.
Sperm yield obtained from SCC was superior to what is usually described in the literature for simple centrifugation (400–600 g), where average losses are 20-25% [20, 31] However, average sperm yields were similar to those obtained by Hoogewijs et al.  for colloidal centrifugation (57.79% vs. 50.9%) protocols. Hoogewijs et al.  reported colloidal centrifugation prior to cryopreservation to be associated with lower sperm yields than standard centrifugation. However, we agree with the theory of these authors that indicate that these lower sperm yields might be offset if colloidally-selected sperm better withstood cryopreservation, thus reducing the sperm dose required to guarantee conception.
In the present work most sperm DNA damage occurs during incubation, not in the first minutes after thawing, in concordance with other authors [6, 9]. This fact remarks the importance of the use of the dynamic form of the SDF assay for evaluating ex vivo procedures . sDFI values were significantly lower immediately after thawing and remained so in the colloidally-selected samples during the first four hours of incubation (24.31 for the SCC protocol vs. 15.59 for Protocol 2). Although there were no significant differences in sDFI values between the SCC protocol and Protocol 1, in the SCC protocol DNA damage occurred more quickly and was more intense. Crespo et al. , López-Fernández et al.  and Cortés-Gutiérrez et al.  indicate that slower DNA fragmentation dynamics may be associated with greater sperm viability in the female genital tract. The results suggest that colloidal centrifugation before cryopreservation would be beneficial in terms of spermatic survival after insemination with frozen-thawed sperm, although further studies are necessary to confirm a possible relation between our “in vitro” observations and “in vivo” fertility.