Trends in sperm quality by computer-assisted sperm analysis of 49,189 men during 2015–2021 in a fertility center from China

Background Sperm quality, including semen volume, sperm count, concentration, and total and progressive motility (collectively, “semen parameters”), has declined in the recent decades. Computer-assisted sperm analysis (CASA) provides sperm kinematic parameters, and the temporal trends of which remain unclear. Our objective is to examine the temporal trend of both semen parameters and kinematic parameters in Shanghai, China, in the recent years. Methods This retrospective study analyzed semen parameters and kinematic parameters of 49,819 men attending our reproductive center by using CASA during 2015–2021. The total sample was divided into two groups: samples that surpassed the WHO guideline (2010) low reference limits (“above reference limit” group, ARL; n = 24,575) and samples that did not (“below reference limit” group, BRL; n = 24,614). One-way analysis of variance, Kruskal–Wallis test, independent samples t-test, and covariance analysis were used to assess the differences among groups. Year, age, and abstinence time were included in the multiple linear regression model of the ARL group to adjust the confounders and depict the trends in sperm quality. Results Among all the total sample and the ARL and BRL groups, the age of subjects increased in recent years. Semen volume and sperm count showed declined tendency with years in the total sample, the ARL and BRL groups, and the subgroup of age or abstinence time, whereas sperm velocities showed increased tendency with years on the contrary. The multiple linear regression model of the ARL group, adjusting for age and abstinence time, confirmed these trends. Semen volume (β1= −0.162; CI: −0.172, −0.152), sperm count (β1= −9.97; CI: −10.813, −9.128), sperm concentration (β1 = −0.535; CI: −0.772, −0.299), motility (β1 = −1.751; CI: −1.830, −1.672), and progressive motility (β1 = −1.12; CI: −0.201, −0.145) decreased with year, whereas curvilinear line velocity (VCL) (β1 = 3.058; CI: 2.912, 3.203), straight line velocity (VSL) (β1 = 2.075; CI: 1.990, 2.161), and average path velocity (VAP) (β1 = 2.305; CI: 2.224, 2.386) increased over time (all p < 0.001). In addition, VCL, VSL, and VAP significantly declined with age and abstinence time. Conclusion The semen parameters declined, whereas the kinematic parameters increased over the recent years. We propose that, although sperm count and motility declined over time, sperm motion velocity increased, suggesting a possible compensatory mechanism of male fertility.


Introduction
Semen quality has been drawing increasing concerns in the recent decades as the fertility rate has declined and infertility problems have become cumulatively serious (1,2). In the late 20th century, sperm concentration and semen volume received the most attention after a systematic review reporting that the two parameters had declined by about 50% over the 50 years from 1930 to 1991 (3). Since then, reports from various regions demonstrated the sperm quality-deteriorated trend including decreased sperm count, concentration, normal morphology rate, and total and progressive motility rates (4)(5)(6)(7)(8)(9)(10)(11)(12)(13). In China, scholars held a similar view, as reported in studies from Hunan (14), Shandong (15), Henan (16), and other regions (17,18). Recently, a systematic review and meta-regression analysis revealed that sperm concentration and total sperm count declined worldwide between 1973 and 2018, and the decline in the 21st century was more rapid than that in the last century (19). However, the decline remains controversial as some studies reported no significant change or even an increase in these parameters (20)(21)(22)(23). Evidently, male fertility decline is destined to be a crucial long-term issue.
Computer-assisted sperm analysis (CASA) is a notable technological advancement that has gradually replaced manual sperm assessment in reproductive centers worldwide. The advantages of CASA lie in its rapid and automatic semen analysis, providing sample statistics and sperm kinematic parameters including curvilinear line velocity (VCL, mm/s), straight line velocity (VSL, mm/s), average path velocity (VAP, mm/s), straightness (STR, %), linearity (LIN, %), the amplitude of lateral head displacement (ALH, mm), beat cross frequency (BCF, Hz), and wobble (WOB, %) (24). These parameters were difficult to acquire by manual analysis. Only a few studies have reported on the clinical implications or trends alterations associated with these parameters (25-28) especially temporal trend of kinematic parameters. The conclusions drawn from CASA parameters or the significance of these parameters need to be evaluated by more investigations and researches.
Sperm count and motility rate as well as kinematic status are pivotal for fertilization. Activated and hyperactivated sperm display specific movement patterns, characterized by high VSL or high VCL and ALH, aimed to propel sperm migration from the cervix to the oviduct and to enhance cumulus cell layers and zona pellucida penetration during fertilization (29-33). VCL  This study explored changes in sperm count, motility, and kinematic parameters in the recent years, analyzing these parameters by year, age, and abstinence time. A comprehensive sperm quality evaluation could provide new ideas on elucidating unexplained male infertility and suggest reference values for the kinematic parameters.

Study population and semen samples
We retrieved data on 83,708 samples assessed at the reproductive center of the International Peace Maternity and Child Health Hospital, Shanghai, China, between January 2015 and July 2021. Subsequently, the samples were screened by the following participants and sample characteristics (1): age, 18-60 (2); abstinence time, 2-7 days; and (3) properly formatted and complete information. Only the first sample report was used when more than one sample from the same individual was available. Finally, 49,189 samples were retained. Although the justifications for sperm detection were not specifically analyzed, most analyses were performed in conjunction with pre-pregnancy clinic and infertility clinic including infertility workup of the patient and his partner, as well as oligo-atheno-spermia and sperm freezing for Assisted Reproductive Technology (ART) cycles. Nearly half of the latter category would have subfertility in the male partner. Because of this, the study population would have been biased toward subfertility and not a random cross section of men of reproductive age.

Ethical approval
This study was under the approval of the Ethics Committee on human subjects of International Peace Maternity and Child Health Hospital (GKLW2018-03).

Semen analysis
The semen samples were collected by masturbation and liquefied at 37°C for at least 30 min. Two well-trained technicians performed all diagnostic semen analyses, and the results were verified by an andrology-trained laboratory director. Sperm volume was evaluated by reading the values directly from the graduated container. Other semen and kinematics parameters were assessed by CASA (Hamilton-Thorne, Beverly, MA, USA). Each sample (5 ml) was loaded into a counting chamber (Leja Products B.V, Nieuw-Vennep, The Netherlands) and analyzed at once. Ten microscopic fields of each chamber were analyzed, evaluating at least 200 spermatozoa. The following parameters were recorded: total count (million), sperm concentration (million/ml), motility (%), progressive motility (%), VAP (mm/s), VSL (mm/s), VCL (mm/s), STR (%), LIN (%), ALH (mm), BCF (Hz), and WOB (%). Sperm were classified as motile when their path velocity exceeded 5 mm/s. Those sperm with path velocities >25 mm/s and linearity > 80% were classified as "progressively motile sperm." The samples were classified as two groups: samples that surpassed the low reference limits ("above reference limit" group, ARL; n = 24,575) and samples that did not ("below reference limit" group, BRL; n = 24,614) according to the World Health Organization laboratory manual for the examination and processing of human semen (fifth edition) (sperm volume, 1.5 ml; sperm count, 39 million; sperm concentration, 15 million/ml; total motility, 40%; and progressive motility, 32%). The samples were also grouped by age (18-24, 25-29, 30-34, 35-39, 40-44, and 45-60 years) and abstinence time (2, 3, 4, 5, 6, and 7 days) to observe trends in these parameters. Semen parameters showed a positively skewed distribution, whereas kinematic parameters were nearly normally distributed. Nonnormally distributed data, semen volume, sperm count, motility, and progressive motility were presented as medians and interquartile range in the tables, and Kruskal-Wallis test was used in the variation analysis, whereas kinematic parameters were shown as means and standard deviation and one-way analysis of variance (ANOVA), independent samples t-test, and covariance analysis were used to assess the differences among groups. Means with confidence intervals (95% CIs) were used in line graphs to depict original trend of parameters. A multiple linear regression model was used to explore the changes in the ARL group parameters controlling for two of three factors: year, age, and abstinence time. The R packages "ggpredict" and "ggplot2" were used to visualize the generalized linear model outcomes. The statistical significance is represented as p < 0.05, p < 0.01, and p < 0.001 or non-significance with p ≥ 0.05.

Distribution of samples
The dataset includes 49,189 semen samples, divided into ARL (n = 24,575, 49.96%) and BRL (n = 24,614, 50.04%) groups, and the samples of each year were shown in Figure 1A. As shown in Figures 1B, C and Table 1, the age of the outpatients differed significantly among years in all groups (all p < 0.001) and showed an increase toward. Abstinence time had significant difference among years with fluctuation toward. The mean age of participants was 32.59 ± 4.50 years, and the mean abstinence time was 4.50 ± 1.70 days in the total sample.

Changes in sperm kinematic parameters
Representative kinematic parameters in the total sample and the ARL group sample over the years were shown in Tables 4, 5, Sample distribution on type, age, and abstinence time. (A) Proportions of the above reference limit group (ARL) and the below reference limit group (BRL). (B, C) Line charts of age (B), and abstinence time (C) based on means with 95% confidence intervals (CIs) over the years. The purple line represents the total sample, the pink line represents the ARL group, and the blue line represents the BRL group.
respectively. The means and 95% CIs of the parameters in the total sample and the ARL and BRL groups were shown as line graphs in Figures  The kinematic parameters of the ARL group-VCL, VSL, VAP, ALH, LIN, and STR-were higher than those of the BRL group (all p < 0.001; Table 6), even after adjusting for age and abstinence time. These parameters could discriminate the sperm motility coincident with semen parameters. Whereas, BCF and WOB of the ARL group were lower than BRL group after the adjustment (p < 0.001).

Multiple linear regression model of ARL group
To exclude the potential effects of discrepant age, abstinence time, and possible diseases in participants of the BRL group on temporal trends in sperm quality, we analyzed the ARL group separately by dividing it into subgroups by age or abstinence time ( We established multiple linear regression models to further explore the relationship between semen parameters with year, age, and abstinence time in the ARL group (Table 7). After adjusting for age and abstinence time, we found significant declines with year in the semen parameters, including semen volume (b1 = −0.162; CI: −0.172, −0.152; Figure    Values are presented as mean ± standard deviation. ARL, above reference limit group; BRL, below reference limit group. Line charts of semen parameters in the three type groups. Means and 95% CIs of semen volume (A), sperm count (B), sperm concentration (C), motility (D), and progressive motility (E) in the groups over the years. The purple line represents the total sample, the pink line represents the ARL group, and the blue line represents the BRL group.   In addition, most semen parameters declined with age (Table 7), including sperm count, semen volume, total motility, and progressive motility, as well as kinematic parameters VCL ( Figure 6A), VSL ( Figure 6C), and VAP ( Figure 6E; all p < 0.001). Sperm concentration and BCF ( Figure 6G, both p < 0.001) increased with age. Moreover, abstinence time was associated with increased semen volume, sperm count, sperm concentration, and BCF ( Figure 6H; all p < 0.001). Conversely, total motility, progressive motility, VCL ( Figure 6B), VSL ( Figure 6D), and VAP ( Figure 6F; all p < 0.001) declined with abstinence time.

Discussion
Male sperm quality decline over the years is a worldwide trend. This study found declining trends over the years in semen volume, sperm count, and motility, consistent with most previous studies (4-18). Unexpectedly, a further novel finding is the increasing trend with year in the kinematics parameters VCL, VSL, and VAP, even after correcting for age and abstinence time. In addition, sperm velocity decreased with age and abstinence time as previous studies have reported (25)(26)(27)(28). This result suggests that continuous attention and improvement of male fertility remain important scientific issues in the field of reproduction.
The cause and clinical significance of the kinematic parameters' increase remain unclear. Sperm velocity is inextricably related to the sperm motility state, which is used to discriminate sperm motile state in CASA and increases during capacitation and hyperactivation (40-42). It was also reported that a higher percentage of hyperactivated spermatozoa was related to higher fertilization rates in conventional IVF (40). Our results hinted that sperm movement velocity increased despite a decrease in the amount and rate of motile sperm, which probably compensates to some extent for male fertility. However, we  cannot exclude the possibility that the activation of sperm motility might be related to premature hyperactivation, which might cause spontaneous acrosome reaction and unexplained fertilization failure (43). This sequence of events has a detrimental effect on male fertility, especially in men with astheno-and oligozoospermia, possibly another indicator of declining male fertility. However, the data of kinematic parameters are not been reported online enough, which made the verification of our results difficult. We hope that more colleagues will pay attention to these data to discuss the accuracy and significance of this result. The causes and clinical implications of these results also should be explored in future prospective and experimental studies.
The age of subjects included in this study increased between 2015 and 2021. The main reason for this increase may be the introduction of the Chinese fertility policy that eased the one-child policy, causing many older men to visit clinical fertility centers. Moreover, there is a tendency to postpone the fatherhood age with the increase in educational requirements for employment. Therefore, we must consider the distribution of population age when studying the trends in semen parameters over time.
Furthermore, age was a significant factor in sperm quality. Many studies have indicated that age was related to sperm concentration, sperm count, total motility, and progressive motility (28, 44). The  Values are presented as mean ± standard deviation. p a , The independent samples t-test of the kinematic parameters between the ARL and BRL groups. p b , Covariance analysis of the kinematic parameters between the ARL and BRL groups, adjusted for age and abstinence time.
kinematic parameters assessed by CASA were also associated with age. Studies showed that ALH, VCL, VSL, VAP, LIN, STR, and head width-to-length ratio significantly decreased with age, suggesting a decline in sperm velocity (25-28). Our results supported this point (Table 7 and Supplementary Figure 1). To exclude the impact of age increase in our study, we compared the semen quality and kinetic parameters of the same age group every year and found that the conclusion that semen quality decreased and sperm motility velocity increased year by year did not change. In addition, age was considered a confounder when analyzing the sperm parameters' temporal trends in regression models. Abstinence time can also impact sperm quality (45-47). Sperm count, volume, and concentration increased gradually with abstinence time of up to 10 days (25, 46). Recent studies focusing on clinical assisted reproduction recommended a short ejaculatory abstinence interval to obtain better sperm quality, especially in infertile men (47- Bar charts of semen quality parameters in subdivided ARL groups by age or abstinence time. The yearly distribution of semen volume, sperm count, VCL, and VAP among the age (A, C, E, G) and abstinence time (B, D, F, H) groups.

52).
Our results showed that semen volume, sperm count, and concentration increased, whereas total and progressive motility, VCL, VSL, and VAP decreased with abstinence time (within the range of 2-7 days). The longer the sperm were stored in the epididymis, the poorer its motility was once ejaculated, especially in the BRL group ( Figure S2). Sperm are continuously stored and naturally renewed in the epididymis. However, many studies suggested that the quality of sperm is closely related to the abstinence time or storage time in the epididymis, especially under pathological or harmful environmental conditions (45,47). First, previous reports demonstrated that human sperm are susceptible to damage induced by reactive oxygen species (ROS) (53). Increased abstinence periods enhance the ROS-exposed time and bring about lipid peroxidation and malondialdehyde that is harmful to acrosome, mitochondrial activity, and DNA integrity (49, 54-58). For subfertile men, abnormal factors may result in excessive production of ROS in the epididymis, leading to decreased sperm motility and an infertile status (59)(60)(61). Impaired antioxidant capacity of abnormal spermatozoa also makes infertile sperm more prone to be damaged by ROS (62-65). Second, longer abstinence period affects secretions of epididymis, prostate, and seminal vesicle, which are related to the pH and biochemical of final semen and affect sperm motility (24, 66, 67). Moreover, longer abstinence time is related to a higher bacterial load and diversity in semen, more pro-inflammatory cytokines (68), as well as decreased rate of morphologically normal sperm (46,69). Therefore, a shorter abstinence time is suggested for patients with oligo-atheno-zoospermia as it increases the probability of obtaining higher-quality sperm.
The rate of decline in sperm count and sperm concentration derived from this study was slightly higher than previous studies (5,18,19) (non-selected population fraction or reproductive center data). This may be related to the demographic and employment characteristics of the city as an economic center. Our research center is located in a mega city of China with high talent competition and socioeconomic burden, and one notable feature that sets it apart from other cities is the high level of work pressure, which, in turn, leads to reduced sleep time, high psychosocial stress, and irregular eating. Studies reported that short sleep durations, late bedtime, and poor sleep quality were associated with impaired sperm health (70,71). Psychosocial stress was also negatively associated with sperm concentration, total count, and motility (72-74). Unbalanced diet is reported among Shanghai population (75) and is harmful to sperm quality, which includes overconsumption of cooking oil and salt as well as under-intake of dairy, vegetables and fruits, soybeans, fish and shrimps, and eggs (75-79). These lifestyles can easily lead to male nutritional disorders and endocrine dysfunction. For instance, overweight and obesity brought by unhealthy lifestyle are positively associated with high estradiol concentrations and lower sperm quality (80, 81). In addition, as China's main industrial center, Shanghai has serious air pollution that adversely impacts on sperm quality (82).
The strengths of this study are as follows. Most previous studies reported that semen parameters declined in recent decades, whereas this study focused on sperm kinematic parameters, showing increasing trends of sperm velocity over the years. Furthermore, we used a multiple linear regression model to analyze the data after    factoring in the effects of age and abstinence time considering that they may have some impacts on the results. Moreover, systematic error in data was low as the methods, personnel, and instruments were consistent throughout the study period. However, the study had several limitations. First, the data came from a single reproductive center, and the subjects included healthy and infertile or potentially infertile men, which might not accurately reflect the trends in the general population. Second, we only corrected for age and abstinence time. Other potentially influencing factors were not recorded or analyzed. Third, the alterations in sperm quality do not necessarily indicate changes in fertility, and the fertilization outcomes were unknown. Despite the shortcomings of this research, the analysis of the results is reasonable and reliable. The majority of sperm quality studies are from singlecenters with minor instrument and manual operation errors. Statisticians afterward could collect many studies' data for systematic review and meta-analysis and summarize widely conclusions. As for this study, the amount of sample is large in this study, and this can offset the impact caused by unselected subjects. We have further subdivided the total sample into different groups to diminish the potential interference. In addition, the conclusion that semen quality is deteriorating year by year is in good agreement with previous reports. Thus, the alterations of sperm kinematics parameters based on mutual reports are analogously convincing in this study. Indeed, semen quality is influenced by factors such as occupational factors, living habits, body mass, and environmental adverse factors. It is extremely difficult to eliminate all factors. Thus, it is crucial to overcome these shortcomings for future prospective experimental research design.
The decrease in semen parameters and increase in the kinematic parameters over the years might be the current trend in male fertility. Sperm kinematic parameters reflect information on sperm movement and functional state. The clinical significance of these alterations needs further studies to be verified and explored, which The plotting of multiple regression linear model of semen and kinematics parameters by year. Images of multiple regression linear models fitting of semen parameters: semen volume (A), sperm count (B), sperm concentration (C), motility (D), and progressive motility (E) and kinematics parameters: VCL (F), VSL (G), VAP (H), and BCF (I) by year adjusted for age and abstinence time.
could provide new reference values for guideline formulation and more diagnosis basis for unexplained infertility.

Conclusion
We presented temporal trends in sperm quality that comprised decreased semen parameters and increased kinematic parameters. Our results implied that enhancive velocity of sperm may be a potential compensatory mechanism of decreased sperm count and motility.

Data availability statement
The data analyzed in this study is subject to the following licenses/restrictions: The dataset belongs to our institution and data underlying this article will be shared on reasonable request to the corresponding authors. Requests to access these datasets should be directed to zhouych@sibcb.ac.cn.

Ethics statement
This study was under approval by the Ethics Committee on human subjects of International Peace Maternity and Child Health Hospital (GKLW2018-03). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Author contributions
YZ, CT, YL, and TL conceived and designed the study. YL and TL completed most of the data collection and analysis. ZMW, ZQW, HW, and TY participated in the sample collection, data analysis, and further explanation. YZ, CT, YL, and TL wrote the manuscript. All authors contributed to the article and approved the submitted version.