Infertility affects almost 15% of couples worldwide who try to conceive (1, 2), and approximately 30% of these couples cannot explain the cause of their infertility (3). Assisted reproductive technology (ART), including in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) has become the most effective treatment for infertility in both developed and developing countries (4, 5). However, improving IVF success rate is still a challenging issue in clinical IVF practice. Despite numerous technical advances in IVF technology, the overall IVF success rate (measured as a live birth)remains low at 16.6-30% (6, 7).

In the past, the success of IVF depended highly on the number of transferred embryos. Increasing the number of transferred embryos may lead to an increased rate of multiple pregnancy, which is one of the most common and serious complications of ART and increases the likelihood of adverse obstetric and perinatal outcomes (8). Currently, the decision of the best embryo for transfer is made based on morphological parameters observed during embryonic development in the IVF laboratory (9, 10). In addition, the widespread use of pre-implantation genetic testing (PGT) has improved the rate of successful pregnancy and reduced the risk of miscarriage (11–13). However, even if the transferred embryo has normal morphological and chromosomal results, some embryos still fail to implant, or the embryos are successfully implanted but fail to result in a live birth (14, 15).

Mitochondria are highly dynamic organelles that play essential roles in the regulation of reproductive processes (16, 17) and intracellular powerhouses that generate cellular energy (adenosine triphosphate, ATP) through the oxidative phosphorylation (OXPHOS) system (18). Human cells harbor hundreds to thousands of mitochondria, each of which contains 2–10 copies of mtDNA in its matrix (19), while in oocytes, there is only one copy per cell (20, 21). Mature oocytes contain more mtDNA than other cells to meet energy demands in the process of fertilization and the early stages of embryonic development (22, 23). Many animal studies have shown that mitochondrial function affects reproductive outcomes (24–26). Recently, an increasing number of studies have related higher mtDNA content in embryos to poor implantation potential and suggested mtDNA content to be a novel tool to predict embryo viability (27–30). Fragouli and Ravichandran et al. established a screening threshold which embryos with mtDNA content above the threshold was less likely to implant (29, 30). Whereas subsequent studies have reached conflicting conclusions (31–33). These contradictory findings limit the application of mtDNA in the accurate assessment of embryo viability in clinical practice. Besides, the most commonly used technology for mtDNA quantification is to calculate the ratio of mtDNA to nuclear DNA (nDNA) in samples using next-generation sequencing (NGS) based or quantitative polymerase chain reaction (qPCR) based methods. In fact, the content of nuclear genomes is not equal across embryos, therefore using solely nuclear genomes for normalization is not enough. To correct for this bias, Victor et al. first proposed the correction factor of embryo’s sex and ploidy for the evaluation of mtDNA content for both NGS-based and qPCR-based technologies and they found no association between mtDNA and implantation potential (34), similar conclusions were later confirmed by another study used the correction factor (35). In addition, El-Damen et al. investigated whether mtDNA content is associated with miscarriage rate and reported that the mtDNA content could not predict the miscarriage of blastocysts after transfer (36).

In general, previous studies have focused more on the relationship between mtDNA and embryo implantation, while few studies have studied the effect of the mtDNA content on live birth. In fact, implantation rate, clinical pregnancy and live birth are commonly used to measure the success of ART treatments. However, for couples who desire a baby, live birth is the only indicator of successful IVF treatment. In addition, patient exclusion criteria have been neglected in some prior studies, which might affect the credibility of the existing results. This study aims to investigate the potential relationship between the mtDNA content and live birth after implantation.

In total, 316 euploid blastocysts from 233 couples were finally included in the study. The relative quantity of mtDNA in each sample was measured using NGS and corrected by a correction factor. We compared the content of mtDNA in blastocysts between different transfer outcomes. The implanted and non-implanted embryos were first analysed, and the epidemiological and clinical characteristics in these two groups were similar except for maternal age (32.1 ± 4.7 vs 33.2 ± 4.7, p=0.05) and biopsy day (p=0.019), data as shown in Table 1 . The relative mtDNA content was 1.01*10⁸ (7.47*10⁷-1.32*10⁸) in the implanted group and 9.91*10⁷(7.08*10⁷-1.40*10⁸) in the non-implanted group. Unexpectedly, no significance was observed between the implanted and non-implanted groups (p=0.965, Figure 1 ).

To assess the relationship between the mtDNA content and live birth after embryo implantation, we then sub-divided the implanted group into the live birth and miscarriage groups, and compared the mtDNA content among the live birth, miscarriage and non-implanted groups, as shown in Figure 2 . The epidemiological and clinical characteristic data are summarized in Table 2 . There was no significance in the mtDNA content among groups (p=0.999, Figure 2 ), blastocysts in miscarriage group did not show a higher mtDNA level than those in the live birth group [median (interquartile range), 1.00*10⁸(7.59*10⁷- 1.39*10⁸) vs 1.01*10⁸ (7.37*10⁷- 1.32*10⁸)]. After controlling for potential confounders in a logistic regression analysis with generalized estimating equations, the relationship between the mtDNA content and live birth remained non-significant (OR=1.000, 95% CI 1.000-1.000, p=0.567), as shown in Table 3 . Further, to determine whether the mtDNA content in day-5 or day-6 embryos was related to embryo viability, we compared the mtDNA content between different transfer outcomes in day-5 and day-6 embryos respectively ( Table 3 ). Similarly, no association between mtDNA content and live birth was observed (p=0.673, p=0.165, respectively).

According to receiver operating characteristic (ROC) curve analysis, the area under curve (AUC) was 0.501 (94.3% sensitivity, 10.9% specificity) for all blastocysts, 0.433 (91.4% sensitivity, 25.0% specificity) for Day 5 biopsied blastocysts and 0.493 (90.0% sensitivity, 18.9% specificity) for Day 6 biopsied blastocysts, which indicates that mtDNA content has a poor predictive value for live birth ( Figure 3 ).

To further evaluate the predictive value of an elevated mtDNA level in reproductive competence, blastocysts with an elevated mtDNA content (above the 95th percentile) were further analysed. Of these 15 blastocysts, four reached live birth, one resulted in miscarriage and the remaining ten blastocysts resulted in non-implanted ( Figure 4 ). The live birth rate among these blastocysts were not significantly different from those among blastocysts with a normal mtDNA content (26.7% vs 33.6%, p=0.780). In addition, it is worth mentioning that the blastocyst containing the highest mtDNA content was from a 44-year-old woman, and the blastocyst resulted in a successful live birth, which suggests that embryos with a higher mtDNA content also have developmental potential for successful pregnancy and live birth.

Maternal age and biopsy day were evaluated for associations with mtDNA content in all 316 embryos. There was no linear correlation between the mtDNA content and maternal age (p=0.570, Figure 5A ) in this study. Further age subgroup analysis also showed no significant difference in the mtDNA content (p=0.997, Figure 5B ) among different age groups (<35, 35-37and ≥38 years). Blastocysts biopsied on Day 5 showed higher levels of mtDNA content compared with those biopsied on Day 6 (p=0.001, Figure 6 ).

Currently, morphological assessment is the primary and most common method for determining which embryo is more suitable for transfer and may yield superior outcomes (52). However, the description of morphological characteristics is subjective, which makes it harder to achieve complete standardization in evaluating the viability and developmental competence of embryos (53, 54). The content of mtDNA, calculated by the ratio of mitochondrial DNA: nuclear DNA (34), is a good objective indicator that has recently been proposed as a new biomarker of embryo viability. Fragouli et al. and Diez-Juan et al. first demonstrated that increased mtDNA levels in embryos were related to poor implantation potential in 2015 (27, 28). Since then, quite a few scholars have conducted research into the relationship of the mtDNA content with embryo viability and transfer outcomes (29–33. 34, 55). Several studies have arrived at the same conclusion (29, 30, 55), whereas others have not (31–33). Victor et al. showed no significant difference in the mtDNA levels between implanted and nonimplanted blastocysts regardless of whether the samples were detected by NGS or qPCR (34). The author also recommended the use of correction factors to normalize the mtDNA content determination among centers, which may reconcile the discrepant conclusions of existing studies. In this study, the calculation of the mtDNA content was corrected by the mathematical factor proposed by Victor et al. (34), and we failed to find any association between the mtDNA content and embryo transfer outcomes (i.e., implantation and live birth). These findings are consistent with those of other studies that also used the same correction factor (35, 56). However, Wang et al. determined the mtDNA content using an NGS-based methodology with mathematical correction but reached the opposite conclusion (57). Wang et al. found that the mtDNA content was significantly associated with embryo implantation and live birth, with lower mtDNA levels leading to higher implantation and live birth rates. One possible explanation is that the range of mtDNA content in human blastocysts varies among centers. In a relatively large, multicenter study by Ravichandran et al. (29), 282 euploid blastocysts from 35 different IVF centers were analysed. The authors found that only half of the centers detected blastocysts with an elevated mtDNA level, and the proportion of blastocysts with a high mtDNA levels did not exceed 27%. Another prospective study from a single large IVF centre showed that 9/199 (4.5%) blastocysts contained high levels of mtDNA (30). In fact, blastocysts with a high mtDNA content seem to be relatively rare, and elevated levels of mtDNA could be associated with culture conditions during laboratory procedures. In addition, the methods used for mtDNA quantification in different studies were not all the same. Thus, the threshold for elevated levels of mtDNA, however, might not be fixed and should be adjusted according to the methods used for mtDNA calculation and culture conditions for embryos in different IVF centers. This may account for some discrepancies among the existing reports. Therefore, the mtDNA content detected in this study, even the highest mtDNA content, may still be in the “normal range” of a euploid blastocyst. It is also understandable that the blastocyst with the highest mtDNA content can achieve a live birth at last.

Mitochondria are the main source of energy in cells, providing more than 95% of the ATP required for cell metabolism (58). The maintenance of a healthy mitochondrial network depends on the proper regulation of mitochondrial quantity and quality. The mitochondrial content is biologically modulated to accelerate ATP generation and facilitate changes in respiratory capacity to keep up with cellular metabolic demands (59, 60). It has been suggested that mtDNA activity and content increase when embryos are under oxidative stress (61, 62). Thus, an elevated mitochondrial content may be a marker of oxidative and metabolic stress in embryos but not a direct marker of mitochondrial dysfunction. Thus, embryos with a higher mtDNA content might also have the potential to develop into a fetus after implantation. Furthermore, according to the quiet embryo hypothesis and Goldilock’s principle (63, 64), the levels of mtDNA in embryos with a better potential for development will be located in a suitable range, called Goldilock’s zone. Embryos lacking sufficient energy reserves usually show a compensatory increase in the mtDNA content to meet the energy requirement during embryo development. Therefore, for embryos with a mild energy deficiency, such a remedy could be very useful and the mtDNA content may still in the Goldilock’s range, the embryo also has the potential for implantation and live birth. Once an embryo is implanted, a higher mtDNA content is less likely to result in adverse outcomes, which is supported by the finding in our study that the mtDNA content was similar in blastocysts resulting in miscarriage and live birth.

Furthermore, endometrial receptivity, embryo viability, and maternal–fetal tolerance are three important determinants of successful embryo implantation (65). Few prior studies have considered the potential confounding effects of maternal factors. In the study, women with autoimmune and endometrial diseases were excluded from this study, and we still found that no relationship observed between different embryo transfer outcomes. Thus, the mtDNA content seems unable to predict embryo implantation or live birth.

Discrepant results have been reported concerning the association of mtDNA with maternal age. Klimczak et al. and Lee et al. reported that the mtDNA content had no correlation with maternal age (33, 35). We failed to find a linear correlation between mtDNA and maternal age, which is consistent with previous studies. After the sub-group analysis of different maternal age groups. Fragouli et al. found a higher mtDNA content in women aged ≥38 years than in those aged <38 years (27), while Treff et al. obtained the opposite results that mtDNA content decreased with advancing age (31). Other studies found no relation with maternal age (28, 34, 57), which is consistent with our finding. Although the hypothesis that reproductive ageing in older women is associated with the content of mtDNA in blastocyst-stage embryos is attractive (27), our results do not support this argument. We also found that blastocysts biopsied on Day 5 showed higher levels of mtDNA content compared with those biopsied on Day 6, which was consistent with previous studies (33, 55, 57, 66). The replication of mtDNA will not be reactivated in the embryo until the blastocyst stage (67), blastocysts biopsied on Day 6 may undergo more cell divisions compared with blastocysts biopsied on Day 5, which may result in a reduction of mtDNA content in Day 6 blastocysts. However, one study by Wu et al. suggest another explanation (66), they proposed a hypothesis claiming that Day 6 blastocysts are developed from those oocytes which contain lower levels of mtDNA content. This means more time and efforts would be required for these embryos to develop into the blastocyst stage. Therefore, the levels of mtDNA content in Day 6 blastocysts are relatively low.

Although the blastocysts biopsied on Day 5 showed a higher potential for implantation and live birth, and Day 5 blastocysts contained higher levels of mtDNA content compared with Day 6 blastocysts. However, this does not imply that elevated mtDNA content is related to implantation or live birth. The level of mtDNA content is just one of the potential factors which could affect the transfer outcomes of Day 5 blastocysts. Other factors, such as embryo quality, may also have an impact on transfer outcomes. Given that the results of logistic regression and ROC curve analysis, we still conservatively considered that mtDNA content is not a reliable screening biomarker for live birth after embryo transfer.

Despite careful selection of the patients, there are some limitations in the present study. First, the sample size for this epidemiological study was relatively small, future studies need to expand the sample size for further generalization. Second, the biopsy samples were collected at a single IVF centre, which may influence the generalizability of our findings despite the consistency of the culture conditions and testing methodologies.

Taken together, the findings of this study indicate that the mtDNA content has poor value for predicting embryo transfer outcomes (i.e., implantation and live birth). The mtDNA content did not differ between blastocysts resulting in miscarriage and live birth. Additionally, the mtDNA content was not correlated with maternal age. We recommend that more stringent inclusion and exclusion criteria and mtDNA quantification standards should be taken into account in subsequent studies.

The data presented in the study are deposited in the SRA repository, accession number PRJNA764530.

The studies involving human participants were reviewed and approved by the ethics committee of the International Peace Maternal and Child Health Hospital. The patients/participants provided their written informed consent to participate in this study.

Conceived and designed the experiments: XZ and SC. Performed the experiments: XL, MY, and WS Analyzed the data: WS. Wrote the paper: XZ. All authors contributed to the article and approved the submitted version.

This work was supported by the National Natural Science Foundation of China (Nos.81971344, 81771638 and 81901495), the research grant from the National Key R&D Program of China (2018YFC1004900) the Shanghai Municipal Commission of Science and Technology Program (No.21Y21901002) Shanghai Municipal Health Commission (GW-10.1-XK07), the Shanghai “Rising Stars of Medical Talent” Youth Development Program Clinical Laboratory Practitioners Program (201972) Shanghai Municipal Commission of Health and family planning (202140110, 20215Y0216).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.