First sperm morphology investigators examined spermatozoa as a whole (head, mid piece, and tail). Moreover, the definition of “normal” potentially fertilizing spermatozoal morphology was derived from observations on spermatozoa recovered from the female reproductive tract [6] and the surface of the zona pellucida [7].

According to the WHO [2010], “for a spermatozoon to be considered normal, both its head and tail must be normal. The head should be smooth, regularly contoured, and generally oval in shape. There should be a well-defined acrosomal region comprising 40-70% of the head area [8]. The acrosomal region should contain no large vacuoles, and not more than two small vacuoles, which should not occupy more than 20% of the sperm head. The post-acrosomal region should not contain any vacuoles. The midpiece should be slender, regular, and about the same length as the sperm head. The major axis of the midpiece should be aligned with the major axis of the sperm head. Residual cytoplasm is considered an anomaly only when in excess, i.e. when it exceeds one-third of the sperm head size [9]. The principal piece should have a uniform calibre along its length, be thinner than the midpiece, and be approximately about ten times the head length. It may be looped back on itself, provided there is no sharp angle indicative of a flagellar break”. Moreover, according to the WHO [2010], an abnormal head morphology should be considered whenever the sperm head is large or small, tapered, pyriform, round, amorphous, vacuolated (more than two vacuoles or >20% of the head area occupied by unstained vacuolar areas), vacuoles in the post-acrosomal region, small or large acrosomal areas (<40% or >70% of the head area), double heads, or any combination of these [10].

MSOME evaluates the morphological state of six subcellular organelles [5]: acrosome, postacrosomal lamina, neck, mitochondria, tail, and nucleus. The first five of these subcellular organelles are considered morphologically normal according to an arbitrary descriptive approach adopted in Bartoov et al. [11] studies examining sperm by transmission and scanning electron microscopy. The criteria for a normally shaped nucleus by MSOME were smooth, symmetric, and oval configurations. An extrusion or invagination of the nuclear chromatin mass was defined as a regional nuclear shape malformation. The nuclear chromatin content was considered abnormal if it contained one or more vacuoles that occupied more than 4% of the normal nuclear area. A sperm cell exhibiting a normal nucleus as well as a normal acrosome, postacrosomal lamina, neck, tail, mitochondria, and no cytoplasmic droplet or cytoplasm around the head was classified as morphologically normal. According to Bartoov et al. [5], the frequency of morphologically normal spermatozoa as defined by the WHO [12] did not correlate to the frequency of morphologically normal spermatozoa as defined by MSOME.

Over the past decade there is an important shift, focusing on the shape of the sperm head and most importantly the presence of vacuoles [13]. The criteria for a normally shaped nucleus by MSOME, were smooth, symmetric, and oval configurations and the nuclear chromatin content was considered abnormal if it contained one or more vacuoles that occupied more than 4% of the normal nuclear area. A sperm cell exhibiting a normal nucleus as well as a normal acrosome, postacrosomal lamina, neck, tail, mitochondria, and no cytoplasmic droplet or cytoplasm around the head was classified as morphologically normal. According to Bartoov et al [5], the frequency of morphologically normal spermatozoa as defined by the WHO [1999] did not correlate to the frequency of morphologically normal spermatozoa as defined by MSOME.

On 2008, Vanderzwalden et al. introduced a more sophisticated approach towards sperm vacuoles by MSOME and their clinical significance. That classification system took a further step forward by focusing not only on single criteria for normal vacuole size as suggested by Bartoov et al. but also focused on vacuoles number. The authors classified sperm vacuoles to: grade I, no vacuoles, grade II, ≤ two small vacuoles (< 4% of head volume), grade III, ≥ 1 large vacuole, and grade IV, large vacuoles with other abnormalities [14]. The authors reported a significant impact on blastocyst formation among 25 patients after sibling oocyte injection - the occurrence of blastocyst formation was 56.3 and 61.4% with grades I and II spermatozoa, respectively, compared with 5.1% with grade III and 0% with grade IV respectively (p < 0.001).

In a consequent classification system Cassuto etal. added sperm head and base to and introduced the formula: (2 X head) + (3 X vacuoles) + (base). Each of the three components is scored as 1 (normal) or 0 (abnormal), as the highest impact is given to vacuoles component with total score range of 0-6. Sperm head assessment is performed according to WHO criteria (described above) and vacuoles evaluation is divided either absent/small (scored 1) vs. large (scored 0) for normal and abnormal, respectively. Accordingly spermatozoa are classified as high- (4-6), medium- (1-3), and low-quality (0). The authors reported not only higher fertilization rate but also improved blastocyst formation by microinjection of higher scored spermatozoa [15].

An interesting study of Pedrix et al. have not only distinguished between ‘large” and “small’ vacuoles but also correlated vacuoles’ appearance by digital imaging system and conventional semen parameters in 440 semen samples. They suggested two different cut-offs: 1) relative vacuole area to sperm head (RVA) of 5.9% which represented as normal threshold with a sensitivity of 0.76 and specificity of 0.56, and 2) RVA occupying 12.4% of abnormal threshold with a specificity of 1 and sensitivity of 0.09 [16]. The advantage of that classification system over others is the objective RVA measurement (opposed to conventional subjective selection [17]) and large sample size which supplied adequate statistical analysis. Furthermore, RVA calculation combines both vacuoles’ number and size and consequently has the potential to be more reliable than relying on vacuoles’ size alone (Table 1).

The lack of consistency and certainty regarding sperm vacuoles limits the value of MSOME in selecting high-quality spermatozoa [18, 19]. Most importantly, vacuoles’ nature, origin, localization, and clinical significance is still a matter of debate, resulting with significant efforts to improve our understanding and establish reliable definition criteria.

The presence of vacuoles with the sperm head has been reported decades ago [20]. However, its significance is still controversial. Since most vacuoles are located in the anterior aspect of the sperm head, their origin have been speculated to derive either from the nucleus or the acrosome. In their important experiment, Gatimel et al. have demonstrated the existence of sperm vacuoles among patients with globozoospermoa - an abnormal sperm shape and function due to the loss of the acrosome - confirming that these substructures derive from the nucleus [21].

Boitrelle et al. [18] used MSOME to select ‘top’ spermatozoa and spermatozoa with a large vacuole but otherwise normal in an attempt to establish a relationship between one type of vacuole morphology and nuclear status. They studied a total of 450 ‘top’ spermatozoa and 450 vacuolated spermatozoa. The rate of non-condensed chromatin was higher for ‘vacuolated’ spermatozoa than for ‘top’ spermatozoa. However, ‘top’ and ‘vacuolated’ spermatozoa did not differ significantly in terms of DNA fragmentation or aneuploidy. In all vacuolated spermatozoa, the acrosome was intact, the plasma membrane was sunken but intact, and the large vacuole was identified as an abnormal, ‘thumbprint’-like nuclear concavity covered by acrosomal and plasmic membrane, suggesting that the large vacuole appears to be a nuclear ‘thumbprint’ linked to failure of chromatin condensation [18]. On the other hand, Kacem et al. [19] have demonstrated that induction of the acrosomal reaction by ionophore significantly increased the percentage of vacuole-free spermatozoa, suggesting that most nuclear vacuoles are of acrosomal origin. Hence, the best spermatozoa selected by MSOME are mostly acrosome-reacted spermatozoa. Thus, IMSI improves IVF outcome by allowing the selection of acrosome-reacted spermatozoa.

MSOME classification development is characterized by increasing focus on sperm vacuoles rather than other sperm organelles and differentiation between small and large (representing normal and abnormal, respectively) vacuoles. However, their characterization remains a prominent challenge mainly due to lack of homogeneous criteria between different studies. On the one hand, several evidences supported their physiological role in sperm production and function and declined their possible association with infertility.

Tanaka et al. investigated the incidence, size, and position of vacuoles in spermatozoa and spermatids from 11 normozoospermic, ten oligozoospermic or asthenozoospermic, four obstructive azoospermic, and three non-obstructive azoospermic men. They reported gradually increasing vacuoles prevalence within the male genital track from 33.7% of spermatids followed by 87.5% in epididymal spermatozoa and 98.3% in the ejaculate [20]. Interestingly, small vacuoles (< 25% of total sperm head) were more frequent in ejaculated spermatozoa opposed to large vacuoles (> 50% of surface area) dominance in testicular spermatids, further emphasizing the differentiation between small and large vacuoles. Montjean et al. demonstrated that acrosome reaction (induced either by hyaluronic acid or follicular fluid) is significantly related to vacuoles appearance within the sperm head, suggesting their physiological role in that process among 35 healthy participants [22]. Moreover, Fortunato et al. have recently declined the negative correlation between vacuoles and live birth, as well as, disorganization of the sperm DNA among 873 men enrolled for assisted reproduction techniques [23]. Importantly, both Montjean et al. [22] and Fortunato et al. [23] used Bartoov et al.’s criteria [4], as one or multiple vacuoles occupied at least 5% of the total sperm head surface (Bartoov et al.’s criteria), whatever their localization.

Most researchers have chosen to focus on large vacuoles [24], assuming these vacuoles are more likely to have pathological significance. Consequently, while data regarding small vacuoles is currently contradictory between normal and pathological significance [25], accumulating evidence support the notion that large vacuoles are related to several DNA abnormalities. During the sophisticated process of spermatogenesis, the DNA of the differentiating sperm undergoes complex nuclear packaging involving the replacement of histones with protamines, leading to highly condensed and protected DNA to ensure safe passage within the male and female genital tracts [26, 27]. The presence of a large sperm-head vacuole was reported to be associated with failure of chromatin condensation and/or DNA damage and/or aberrant chromosome numbers in spermatozoa [28-30].

Various studies demonstrated abnormal sperm chromatic compaction in spermatozoa with large vacuoles compared to normal spermatozoa assessed by MSOME. For example, Franco et al. described that 67.9% of spermatozoa with large vacuoles (defined as > 50% of sperm head area) had abnormal DNA compaction compared to 33.1% of normal spermatozoa [28]. Similarly, Boitrelle et al. reported 36.2% vs.7 .6% abnormal compaction, respectively, using large vacuoles cutoff as > 25% of sperm head area. Comparison between sperm with large vacuoles and unselected spermatozoa demonstrated identical findings [18]. Similarly, Pedrix et al. assessed by aniline blue staining compared among 20 patients with teratospermia. They reported significantly altered chromatin condensation in spermatozoa with large vacuoles (defined as > 13% of sperm head area) compared to normal spermatozoa [30]. Therefore, in spite of the variable criteria for large vacuoles definition, accumulating evidence support the notion that these features are related to non-efficient chromatin compaction.

Disorganization of sperm chromatin exposes it to damage through various mechanisms such as enzymatically induced DNA breaks, radical oxidants species, and others [31], resulting with DNA fragmentation and male infertility [26]. Recently, Pastuszek et al. examined DNA fragmentation among almost 4,000 spermatozoa from ten patients selected by MSOME. They demonstrated that abnormal Vanderzwalden classification was associated with increased DNA fragmentation percentage, especially resulting from double-strand DNA breaks [32]. Hammoud et al. classified spermatozoa from eight patients according to motility and morphology alone by either ×200 or ×6,300 magnifications (representing ICSI and IMSI, respectively), and reported significantly lower DNA fragmentation among normal spermatozoa selected by MSOME [33]. Increased DNA fragmentation was reported as well among spermatozoa in with large vacuoles defined either by 5-50% or > 50% of sperm head area [28, 34]. More specifically, Utsuno et al. demonstrated statistically significantly higher percentage of DNA fragmentation in spermatozoa with abnormal ellipticity and abnormal angularity than in spermatozoa with normal-shaped heads (6.1% and 5.4% vs. 2.8%) on MSOME. Spermatozoa with large nuclear vacuoles also correlated with sperm DNA fragmentation, and had a statistically significantly higher percentage of DNA fragmentation (4.7%) [35]. In their review, Boitrelle et al. emphasized that DNA fragmentation association with spermatozoa vacuoles is mainly relevant in semen samples with overall high DNA fragmentation [36].

In light of these multiple and repetitive findings, it seems that large vacuoles are associated with impaired of chromatin organization and increased DNA fragmentation. However, one of the few studies which focused on small vacuoles demonstrated DNA disorganization in cases of multiple compared to single vacuole [25]. Therefore, there is a possibility that both vacuoles’ size and number are related to DNA in efficient packaging.

In addition to DNA fragmentation, several researchers evaluated the possible correlation between sperm selection methodology and sperm aneuploidy. Levron et al. reported that motility-based selection by ICSI among eight patients with male factor and repeated implantation failure resulted with significantly lower sperm chromosomes X, Y and 18 aneuploidy compared to morphology-based selection [37]. Garolla et al. performed CGH for individual spermatozoa with either large vacuoles, small vacuoles, or no vacuoles. Sperm selected from infertile patients showed a significantly higher percentage of abnormal molecular karyotypes than controls (19.4 vs. 7.7%, respectively). Interestingly, sperm with large vacuoles and small vacuoles showed 38.3% and 20.0% abnormal karyotype in infertile men versus 18.3 and 5.0% in controls, respectively (p < 0.01). Moreover, complex abnormalities were found only in the large vacuoles category. Abnormal karyotype was never found in normal sperm [38]. Although that study included only three patients in each group, it confirmed the association between vacuoles and DNA abnormalities and further emphasizes the pathological significance of the large vacuoles. Pedrix et al. reported similar results with significantly higher aneuploidy and diploidy rates among semen samples with large sperm vacuoles compared to controls [30]. In spite of these findings, these results are not sufficient to confirm an absolute impact on embryo aneuploidy. The present authors have recently reported that abnormal DFI (a possible reflection of abnormal sperm vacuoles) had no correlation with blastocyst aneuploidy examined by PGS-CGH after ICSI [26, 27]. Although PGS limitations have been reported exclusively over the last few years [39, 40], similar study design which will investigate embryo euploidy after IMSI may supply important data on that important subject.

In conclusion, multiple and various studies demonstrates convincing data regarding the association between large sperm vacuoles and DNA abnormalities. Utsuno et al. compared 2,400 normal spermatozoa with 2,400 protamine-deficient spermatozoa among 36 infertile men and reported that protamine deficiency is related not only to abnormal sperm head morphology, but also associated with higher DNA fragmentation and large nuclear vacuoles’ appearance [41]. It seems that inadequate chromatin compaction leads to higher rate of DNA fragmentation and aneuploidies, which are followed by the appearance of the large vacuoles. On the other hand, small sperm vacuoles are probably physiological phenomenon and may be related to sperm maturation within the male genital track. Distinguishing between large and small vacuoles remains a challenge, and clear cutoff criteria are further to be established and confirmed. In light of that inconsistency, clinical studies have been conducted in efforts to evaluate the clinical significance.

Various studies have attempted to implement IMSI into clinical usage instead of the gold standard ICSI procedure. While some prospective studies were performed (most of them more than five years ago), most clinical researches had a retrospective design. Overall, these studies are characterized by very wide heterogeneity, specifically with regards to inclusion criteria in which IMSI may be used. A recent Cochrane review comparing the effectiveness and safety of IMSI and ICSI in couples undergoing ART [42] found no effect on live birth or miscarriage nor evidence that IMSI improves clinical pregnancy. They concluded that the results from RCTs do not support the clinical use of IMSI and further trials are necessary to improve the evidence quality before recommending IMSI in clinical practice. However, these results may be attributed to different patient populations and MSOME classification included in the meta-analysis. Although numerous retrospective trials have been investigated the impact of IMSI within the clinical settings, for the purpose of the current review, the present authors decided to focus on prospective studies which obviously supply a higher quality of scientific evidence. Most importantly, substantial egg factor was excluded by either excluding female patients above age of 40 (and even younger age in others) or by including patients with at least 6-8 retrieved eggs. However, as presented in Table 2, these studies lack uniform methodologies such as inclusion criteria, comparison between study and control groups, and MSOME classification criteria.

Based on the above observations, it is not surprising that abnormal sperm vacuoles, related mainly to sperm DNA abnormalities, have been suggested to have detrimental impact on human reproduction. Specific attention was focused on differentiating between early and late paternal affects, which are attributed to oocyte activation vs. paternal DNA activation on day 3 after fertilization, respectively [43].

Vanderzwalmen et al. [14] reported significantly higher blastocyst formation among couples with grades I-II MSOME classification (56.3%-61.4%) compared to grade III and IV (5.1% and 0%, respectively, p < 0.0001, described above). In their prospective study, Setti et al. performed MSOME immediately after ICSI among couples undergoing first IVF-ICSI cycles due to male factor. They reported that large vacuoles (defined as > 13% of sperm head area) were mainly associated with impaired blastocyst formation including lower rates of normal inner cell mass and trophoectoderm development. Additionally, they demonstrated reduced number of blastomeres in earlier days 2 and 3. Importantly, these developmental modifications were not related to conventional sperm morphology assessment [44]. In their randomized prospective study, Knez et al. included infertile couples with abnormal semen analysis in whom all embryos in previous ICSI cycle were arrested prior to blastocyst stage. Couples were divided to IMSI group (n=20 using the Cassuto et al. classification system) vs. repetitive ICSI (n=37). The authors reported significantly lower rate of repetitive arrested embryos in the IMSI compared to ICSI group (0% vs. 27%, p < 0.05). While clinical pregnancy rate in the IMSI group was higher as well (25% vs. 8.1%), it did not reach statistical significance [45]. Therefore, both Vanderzwalmen et al.’s [14] and Cassuto el al.’s criteria system support the notion that IMSI may improve blastocyst formation by improving the late paternal affect after the activation of the embryonic genome.

The first report of Bartoov et al. [4], which included couples after five ICSI failures, pointed that clinical scenario as a potential to benefit from IMSI. Retrospective studies are generally characterized by different methodologies and therefore their scientific significance is limited. For example, Shalom-Paz et al. reported significantly higher pregnancy rate and reduced spontaneous abortions using IMSI compared to previous ICSI after three IVF-ICSI failures [46]. On the other hand, Gatinel et al., while assessing the benefit of IMSI in patients undergoing their third ART attempt, IMSI did not improve the clinical outcomes compared with ICSI, according neither to implantation, clinical pregnancy, nor to live birth rates [47].

Most importantly, few prospective randomized trials have been performed. Antinori et al. performed one of the largest randomized studies focused on couples with isolated male factor in their first-third IVF cycle, including 219 couples that were treated by ICSI and compared to 227 couples treated by Bartoov et al.’s IMSI criteria using motile spermatozoa with normal head dimensions (length 4.75 ± 0.28 μm, width 3.28 ± 0.20 μm) and shape, and with no or a maximum of one vacuole (0.78 ± 0.18 μm). They reported that IMSI vs. ICSI resulted in a higher clinical pregnancy rate (39.2% vs. 26.5%, p = 0.004) and lower miscarriage rates (17.4% vs. 37.5%). Most importantly, a subgroup analysis demonstrated that IMSI clinical pregnancy rate superiority was isolated among patients after two previous cycles (29.9% vs. 12.9%, p = 0.017), while no significant differences were found between IMSI and ICSI in first and second cycles [48]. In another prospective study in which couples acted as their own controls, 75 infertile couples were offered IMSI after at least two previous IVF or ICSI failures. Semen parameters were not related to inclusion criteria. The researchers used Bartoov et al.’s criteria for spermatozoa selection for IMSI. The authors reported improved day 2 and blastocyst formation in IMSI compared to IVF/ICSI cycles (89.8% vs. 79.8%, p = 0.009 and 41.3% vs. 26.7%, p = 0.04, respectively), as well as higher clinical pregnancy and birth rates of 29.3% and 18.6%, respectively [49]. A similar study design, which included eight couples after single ICSI failure who had identical ovarian stimulation protocol and IMSI in their subsequent cycle, resulted with significantly improved grade A embryos formation (83.6% vs. 60.3%) and implantation rate (20.8% vs. 0%) in the subsequent IMSI cycle, resulting with three pregnancies and five live births. It should be noted that the authors selection of MSOME criteria included poor morphology, presence of multiple vacuoles, presence of vacuoles over 4% of area, and poor morphology of midpiece [50]. Opposite to these reports, El Khattabi et al. reported a similar implantation rate (16.1% vs.16.7%, p = 0.77) and live birth rate (21% vs. 22%, p > 0.99) between IMSI by Vanderzwalmen et al.’s criteria (n=90) and ICSI (n=130) among couples with mild male factor who had at least two previous implantation failures after transfers of good-quality embryos [51].

In conclusion, there are is evidence which supports recurrent IVF/ICSI failures with male factor as a possible indication for IMSI. However, data is inconsistent between studies, possibly due to various spermatozoa selection criteria during MSOME and the definition of repeated IVF failure (cycle rank). Further studies with homogenous methodologies should be performed prior to routine applicability.

De vos et al. conducted a prominent research, during which sibling oocytes were randomized to IMSI vs. ICSI in couples with male infertility (> 0.1×10⁶ spermatozoa/ml) without female factor. Out of 350 cycles, 188 were first ART cycle while 72, 50, and were 27 second, third, and fourth cycles, respectively. The researchers used Vanderzwalmen et al.’s criteria for IMSI. In addition to similar fertilization rate (79.1% and 77.3%, respectively, p = 0.22), the authors reported similar preimplantation development up to blastocyst stage and clinical outcome although embryo transfer was not randomized [52]. The researchers should be acknowledged for the strict methodology and large sample size of 350 cycles which supply a high level of scientific reliability to their findings. Although the study included patients with various ART attempts, the present authors believe the results mostly reflect first cycle (53.7% of total cases). Another prospective randomized trial included couples with male factor (at least three million of spermatozoa in the ejaculate and less than one million of motile spermatozoa recovered after density gradient with whatever sperm morphology) during first ART and compared IMSI (n=116) using Pedrix et al.’s criteria (described above) vs. conventional ICSI (n=139). No statistically significant difference was observed between the two groups with regard to the rates of implantation (IMSI: 24%; ICSI: 22%; NS), ongoing pregnancy (IMSI: 31%: ICSI: 33%, NS) and birth (IMSI: 27% ICSI: 30%; NS). Interestingly, the fertilization rate was significantly lower in the IMSI group (56 ± 25 vs. 63 ± 23, p < 0.05), as well as the total number of embryos (4.8 ± 3.2 vs. 5.8 ± 3.9; p < 0.05), and the number of frozen embryos (1.4 ± 2.3 vs. 2.2 ± 3.0; p < 0.05). The authors speculated that the technical constraints of IMSI (time) affect the quality of gametes. However, such a deleterious effect of IMSI has not been reported in other studies [53]. Taken together, these prospective studies decline isolated male factor as an indication for IMSI in first ART cycle. On the other hand, Wilding et al. performed a randomized controlled trial comparing between ICSI and IMSI (125 patients in each group) among couples with male factor (sperm count 1-20 million per ml) without describing ART cycles history. Selection MSOME criteria were poor morphology, presence of multiple vacuoles, presence of vacuoles over 4% of area, and poor morphology of midpiece. The rate of fertilization was not significantly different between the cohorts. The authors reported a significantly higher rate of grade I embryos in the IMSI group compared to conventional ICSI (98.6% and 66.0% of transferred embryos, respectively, p < 0.05). The pregnancy rate of patients undergoing IMSI was significantly increased with respect to the ICSI controls (65.6% vs. 40.0%, p < 0.05). The implantation rate of embryos created with IMSI procedures was greater than that of standard ICSI techniques [50]. Unfortunately, the lack of ART cycle number is a major drawback of the study especially as the current literature differentiates IMSI implementation between couples undergoing their first cycle and those with previous ICSI failures.

An interesting optional indication for IMSI is male factor with isolated teratospermia, since the main theoretical advantage of MSOME is detailed morphological evaluation. In their prospective randomized study, Knez et al. reported significantly higher clinical pregnancy rate in the IMSI group (n=52 using Vanderzwalmen et al.’s criteria) compared with ICSI group (n=70, 48% vs. 24%, respectively, p < 0.05). Couples in the IMSI group had an average of 2.6 previous ART cycles compared to 2.1 in the ICSI group. After IMSI, a statistically higher number of morulae developed and a lower number of embryos arrested at low-cell developmental stages compared to ICSI (21% vs. 13%, p < 0.05; and 44% vs. 62%, p < 0.01, respectively). In a further analysis the authors demonstrated significantly higher blastocyst formation using spermatozoa ranked I and/or II compared to III and/or IV, supporting previously report of Vanderzwalmen et al. [54]. El Khattabi et al. compared IMSI (using Vanderzwalmen et al.’s criteria) and ICSI among patients with teratospermia in their first-second ART cycle. The number of fertilized oocytes (5.2 vs. 4.3, p = 0.029) and cleaved embryos (5.3 vs. 4.4, p = 0.03) were significantly higher in the IMSI subgroup (n=132). Additionally, a significantly higher number of top-quality embryos (2.3 vs. 1.7 [p = 0.009] at day 2; 1.7 vs. 1.1 [p = 0.003] at day 3, and good-quality embryos 3.1 vs. 2.4 [p = 0.026] at day 2, 3.3 vs. 2.3, [p = 0.019] at day 3) was observed after IMSI compared with the ICSI subgroup (n=126). Most importantly, an implantation rate (30.7 vs. 20.1%, p=0.007) and live birth rate (38% vs. 20%, p = 0.002) were significantly higher after IMSI compared with ICSI [51]. Taken together, both studies supply convincing data regarding the effectiveness of IMSI using Vanderzwalmen et al.’s criteria for couples with isolated teratospermia. However, it should be noted that in Knez et al. study, the average of previous ART cycles were 2.6 and 2.1 in the IMSI and ICSI groups, respectively. Therefore IMSI superiority may be explained not only by the teratospermia selection but alternatively by previous failures.

The rationale for MSOME as improved normal spermatozoa detector compared to ICSI has led few researchers to prospectively investigate IMSI outcome in first ART cycles among unselected infertile patients. Balaban et al. failed to demonstrate significant improvement in the clinical outcome using IMSI using Bartoov et al.’s MSOME criteria. However, specifically severe male factor patients in the IMSI group had significantly higher implantation rates compared with their counterparts in the ICSI group (29.6% vs. 15.2%, p = 0.01). Unfortunately, number of ART cycle was not reported [55]. The lack of efficiency of MSOME among unselected patients population is further supported by Gatimel et al., who found no sperm head vacuoles differences between fertile and idiopathic infertile patients [56]. In other words, MSOME and IMSI have no diagnostic or therapeutic added values among non-selected or idiopathic infertile patients, on the contrary to couples with male factor in whom these methodologies may be offered under specific indications.

The advantage of IMSI to reduce spontaneous abortions due to declining spermatozoa with DNA abnormalities has been previously suggested [48, 57]. Although theoretically relevant, the correlation between DNA fragmentation (a possible trigger for abnormal vacuoles formation) and spontaneous abortions is a matter of debate [58]. Therefore the notion that IMSI by itself may reduce the risk for spontaneous abortions should be validated by prospective research focused on patients with recurrent pregnancy loss. To the best of the present authors’ knowledge, such a study has not been conducted yet.