Intra-individual changes in sperm DNA fragmentation levels over short and long time periods

Abstract

Purpose

We aimed to examine the longitudinal, intra-personal changes in DNA fragmentation index (DFI) over time.

Methods

Men who performed at least two DFI measurements (using sperm chromatin structure assay (SCSA) between 2003 and 2019 were included in this study and allocated to groups by time between DFI tests: < 1 year, 1–3 years, 3–5 years, and > 5 years. An analysis of DFI change over time according to age groups was additionally performed. Regression models were developed to predict changes in DFI with time.

Results

Overall, 225 patients had two or more DFI measurements done at least a month apart (mean of 586.7± 710.0 days). The < 1 year (n = 124) and 1–3 years (n = 68) groups demonstrated decreased DFI levels, while an increase in DFI was shown in 3–5 years (n = 21) and more than 5 years (n = 12) groups − 7.1 ± 14.9%, − 4.5 ± 13.4%, + 3.2 ± 8.4%, and + 10.8 ± 18.0%, respectively, p < 0.001). This trend was similarly shown in age subgroups of under 40 years and 40–50 years at baseline DFI. Linear regression models showed that the factors predictive of DFI increase are baseline DFI and > 3 years between DFI tests.

Conclusion

This study shows that DFI, in men being investigated for infertility, initially decreases in the first 3 years of follow-up, and then increases over time with the highest increase occurring after 5 years interval (an average increase of 10.8%). Testing infertile men’s DFI levels at first evaluation may contribute to personalized consult regarding future reproductive outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-023-02891-z.

Introduction

Infertility is estimated to impact nearly 15% of all couples worldwide, with male infertility etiology contributing to approximately 40–50% of cases [1–3].

The global trend of delayed fatherhood over the last few decades [4, 5] gave rise to increasing interest in the effect of advanced paternal age on sperm quality and reproductive outcomes [6, 7]. Sperm quality is affected by many factors; among them is paternal age. Advanced paternal age was shown to be associated with higher rates of abnormal sperm parameters, especially decreased progressive motility and higher DNA fragmentation levels [8–11].

The evaluation of DNA integrity by DNA fragmentation index (DFI) in the context of male infertility has gained popularity over the last decade and is now commonly used for semen quality assessment in infertile men with normal and abnormal sperm parameters as well as in cases of recurrent miscarriages [12, 13]. Moreover, sperm chromatin integrity and the measurement of DFI have been recently recommended in the 6th edition of the WHO guidelines [14], as it plays an important role in the evaluation of sperm quality alongside the conventional semen analysis parameters. Several factors were previously shown to increase sperm DNA fragmentation (SDF). Among these factors are obesity, smoking, alcohol consumption, and exposure to environmental factors that increase oxidative stress [15].

The importance of DFI in the context of assisted reproductive technology (ART) outcomes has been thoroughly investigated. Fertilization rates, embryo quality, clinical pregnancy rates, and spontaneous abortion rates were all shown to correlate with DFI levels 1[6–19], though some studies demonstrated that increased DFI does not alter ART outcomes [20–23].

While the association between advanced paternal age and DFI levels was previously shown in cross-sectional studies showing higher DFI in older men [24–27], the intrapersonal change in DFI over long period of time, which involved longer exposure to negative factors, was not previously assessed.

Studies on changes in individuals have the potential to enable development of useful predictive models. Therefore, in this study, we aimed to investigate the intrapersonal changes in DFI over time in a population of infertile men, which is the population of interest, and evaluate whether the age at baseline DFI measurement contributes to this change.

Methods

Study population

We performed a retrospective cohort study at a Canadian university–affiliated infertility clinic between 2003 and 2019 and included infertile men who were referred for infertility evaluation and provided at least two semen samples tested for sperm DNA fragmentation 1 month or more apart. Exclusion criteria were men who presented with previous exposure to risk factors known to affect semen quality—chemotherapy, radiotherapy, testicular or other cancer diagnosis, men diagnosed with Klinefelter syndrome, and those who underwent orchiectomy. All men included in the study had both of their samples assessed in the same sperm laboratory (Mount-Sinai Hospital, Toronto, Canada) for general semen analysis and DNA fragmentation index evaluation, and the same technique for DFI assessment was used throughout the entire study period.

Men included in the study were allocated to four groups according to time between DFI tests—(1) less than 1 year, (2) 1–3 years, (3) 3–5 years, and (4) more than 5 years. Age subgroups groups were also analyzed, comparing three categories—under 40 years, 40–50 years, and older than 50 years.

Primary outcome was the change over time, per patient, between the baseline DFI and last DFI measured.

Semen analysis and data collection

Semen samples were provided by means of masturbation and analyzed according to the 2010 WHO criteria [28] for normal abnormal sperm parameters—semen volume (milliliter (ml)), sperm concentration (million/ml), overall and progressive motility (%), sperm morphology (%), and viability (%).

An analysis of two specimens from the semen sample was used to determine the average value of sperm concentration and motility. This was done by the Makler chamber (Sefi Medical Instruments, Israel) and computer-assisted sperm analysis (CASA). Morphology assessment was done using a microscopic high-power evaluation (at least 200 sperms).

Viability was measured by eosin-nigrosin assay, as previously described [29], and viable (sperm head unstained) and non-viable (sperm head stained) sperm counts were recorded.

As previously described [30, 31], DNA fragmentation index was calculated by testing frozen prepared semen sample using the sperm chromatin structure assay (SCSA). Briefly, samples were treated for 30 s with 400 μL of a solution of 0.1% Triton X-100, 0.15 M NaCl, and 0.08 N HCl. Thirty seconds later, staining buffer (1.2 mL) was admixed to the test tube, and flow cytometry activated cell sorter (Caliburflow cytometer, Becton Dickinson, San Jose, California) was used for the analysis of a minimum of 5000 cells. The percentage of sperm with denatured DNA, reflected by the proportion of cells exhibiting an abnormal emission of red fluorescence, was recorded. All men attending the infertility clinic for initial infertility assessment filled out a computer-based survey after signing the IRB-approved informed consent form. This survey included data regarding baseline demographics, medical history, and fertility-related information at first visit to the infertility clinic—age, body mass index (BMI), smoking status, alcohol and cannabis use, diabetes, medication use (including fertility supplements), duration, and type of infertility (primary vs. secondary). Semen analysis and DFI level data from the Mount-Sinai Hospital Andrology laboratory were later linked with the abovementioned information.

Statistical analysis

Patient characteristics were summarized using descriptive statistics. Differences in characteristics by timing between first and last DFI measurements were assessed using Kruskal-Wallis tests and Fisher’s exact tests for continuous and categorical characteristics, respectively. Post hoc pairwise comparisons for statistically significant associations were conducted using Dunn’s test and Fisher’s exact tests with the Holm correction for continuous and categorical characteristics, respectively. Change in DFI within each group was assessed using the signed-rank test.

Scatterplots and loess curves were used to visualize the relationship between age and DFI change, and between measurement timing and DFI change. The relationships were also summarized using Spearman’s correlation coefficient.

Linear regression models were used to examine the association between patient characteristics and time between DFI measurements and change in DFI, and logistic regression models were used to examine their association with the odds of a 5% and 20% in DFI. For all outcomes, univariable regression models were first conducted, and all characteristics that were significant in the univariable analysis were retained in the multivariable regression model. Statistical analyses were performed in R, version 4.2.0. All comparisons were two-tailed and, p values < 0.05 were considered statistically significant.

Ethical approval

All participants have signed the IRB-approved informed consent form. The Research Ethics Board of the Mount Sinai Hospital approved data collection (reference number 05-0161-E) and analysis (reference number 07-0032-E). The dates of approval were October 18, 2005, and October 30, 2007, respectively.

Results

Overall, 225 men with two or more DFI measurements done at least 1 month apart between 2003 and 2019 were included in the study. The mean age at baseline DFI evaluation was 38.9 ± 5.8 years and the average baseline DFI value was 40.3 ± 17.3%. The average infertility period was 3.7 ± 3.7 years, and 19/225 (8.4%) of men were referred due to secondary infertility.

In the entire study population, the mean time between the first and last DFI tests was 586.7 ± 710.0 days (median of 297 days, range 1 month–12 years). The proportion of patients with a timing between tests of < 1 year, 1–3 years, 3–5 years, and > 5 years were, respectively, 55.1%, 30.3%, 9.3%, and 5.3%.

Basic characteristics and first semen analysis data, including DFI, were compared across the four time groups (Table 1). There were no differences observed in demographic and medical background. Five men in the < 1 year group (4.0%), five men in the 1–3 years group (7.4%), and zero in the 3–5 years and > 5 years groups were taking multivitamins or fertility supplements when performing their first DFI test (NS). Semen parameters, including volume, concentration, motility, and morphology, were similar across groups. Baseline DFI levels were 42.9 ± 16.0%, 40.4 ± 1.3%, 28.0 ± 16.3%, and 37.7 ± 16.6% (p = 0.001), respectively, in the four groups (Table 2). Post hoc pairwise comparisons indicated that the differences between < 1 year vs. 3–5 year groups and the 1–3 years vs. 3–5 years groups were statistically significant (p < 0.05). There were no differences in last DFI value across groups. The intrapersonal change in DFI (between the baseline and last DFI) was assessed for each patient.

Table 1.

Characteristics of the study population—infertile men with two DNA fragmentation tests according to time between the first and last test

Parameter	Time between first and last DFI measurements
	< 1 year	1–3 years	3–5 years	> 5 years	P valuea
Number of patients	124	68	21	12
Age (years)	38.9±5.7	38.7±5.9	39.8±6.6	37.8±5.0	0.85
BMI	26.9±2.9	27.0±3.1	N/A	24.1±0.4	0.41
Missing	93	58	20	10
Smoker	2 (1.8%)	4 (6.6%)	0 (0%)	0 (0%)	0.38
Missing	7	7	1	1
Alcohol use	73 (58.9%)	34 (50.0%)	10 (47.6%)	3 (25.0%)	0.11
Marijuana use	4 (3.2%)	0 (0%)	1 (4.8%)	0 (0%)	0.30
Diabetes	4 (3.2%)	2 (2.9%)	0 (0%)	0 (0%)	1.00
Years of infertility	3.3±3.6	3.5±2.9	3.8±3.1	8.2±6.7	0.05*
Primary infertility	87 (86.1%)	59 (93.7%)	20 (95.2%)	9 (100.0%)	0.34
Missing	23	5	0	3
Medications use	32 (25.8%)	24 (35.3%)	6 (28.6%)	4 (33.3%)	0.55
Baseline DFI (%)	42.9±16.0	40.4±18.3	28.0±16.3	34.7±16.6	0.001*
Baseline semen parametersa
Volume (ml)	2.8±1.4	2.9±1.6	3.3±1.6	2.8±1.0	0.81
Concentration (million/ml)	37.4±43.0	46.6±60.0	40.0±32.9	86.6±79.4	0.41
Motility (%)	18.1±13.1	19.8±13.2	22.9±13.9	14.4±8.8	0.45
Normal morphology (%)	11.6±10.0	15.7±11.1	16.7±10.3	12.0±15.7	0.06
Missing	18	14	7	10

Data presented as n(%) or mean ± SD

DFI DNA fragmentation index, BMI body mass index

^aKruskal-Wallis test for continuous parameters; chi-squared test for categorical parameters * Statistically significant (p < 0.05)

Table 2.

Comparison between the baseline and last sperm DNA fragmentation tests according to time between tests

Parameter	< 1 year	1–3 years	3–5 years	> 5 years	P valuea
Number of patients	124	68	21	12
Baseline DFI (%)	42.9±16.0	40.4±18.3	28.0±16.3	34.7±16.6	0.001*
Last DFI (%)	35.8±17.2	35.8±17.3	31.1±17.6	45.5±23.8	0.28
DFI absolute difference per patient	-7.1±14.9	-4.5±13.4	3.2±8.4	10.8±18.0	< 0.001*
DFI relative change per patient	-13.7±31.3	-3.9±37.0	17.0±36.5	41.4±71.6	< 0.001*
DFI Difference Category					0.001*
Decrease (less than − 2.0%)	71 (57.3%)	37 (54.4%)	4 (19.0%)	3 (25.0%)
No Change (greater than or equal to − 2.0% and ≤ 2.0%)	22 (17.1%)	8 (11.8%)	3 (14.3%)	1 (8.3%)
Increase (>2.0%)	31 (25.0%)	23 (33.8%)	14 (66.7%)	8 (66.7%)

Data presented as mean ± SD or count (%)

DFI DNA fragmentation index

^aKruskal-Wallis test for continuous parameters, Fisher’s exact test for categorical parameters * Statistically significant (p < 0.05)

Change in DFI from baseline to last measure in the < 1 year (mean − 7.1 ± 14.9%) and 1–3 years (− 4.5 ± 13.4%) was statistically significant (signed-rank p < 0.001 and p = 0.017, respectively; Table 2). The changes in the 3–5 years (3.2 ± 8.4%) and 5+ years (10.8 ± 18.0%) were not statistically significant (signed-rank p = 0.065 and p = 0.077, respectively). A comparison of the absolute and relevant change between the time groups was statistically significant (p < 0.001). Post hoc pairwise comparisons indicated that the differences between < 1 year vs. 3–5 year groups the 1–3 years vs. 3–5 years groups and the > 5 years vs. < 1 year and 1–3 years groups were statistically significant (p < 0.05).

Among men with repeated DFI within 3 years (n = 192), 108 (56.3%) had improved DFI levels, 30 (15.6%) had the same DFI levels (change of 2 percentage points or less), and 54 (28.1%) had higher DFI. Amongst the patients with an increase in DFI, the mean increase was 8.5%. Amongst men who had their last DFI more than 3 years after their first (n = 33), 7 (21.2%) had lower DFI, 4 (12.1%) had similar DFI, and 22 (66.7%) had higher DFI. Amongst men with an increase in DFI, the mean increase was 12.2%.

The rate of men with baseline DFI under 30% who had their last DFI exceeding 30%, was 12.5% (2/16) in the < 1 year group, 11.8% (2/17) in the 1–3 years group, 16.7% (2/12) in the 3–5 years group, and 50% (2/4) in the group of patients with more than 5 years between DFIs (NS).

DFI changes by timing between tests for three age subgroups (under 40 years, 40–50 years, and above 50 years) was analyzed (Table 3). The 3–5 years and 5+ years groups were collapsed due to the small number of patients in the latter group. Differences in DFI change were observed amongst patients aged ≤ 40 years (p = 0.003 and p = 0.004 for change and percent change, respectively) and amongst patients aged 41–50 years (p = 0.002 and p < 0.001). Post hoc pairwise comparisons indicated that for both age groups, DFI change differed across the < 1 year and 1–3 years groups, and < 1year vs. > 3 years groups). In Fig. 1, the relationship between number of days between DFI tests and DFI change was visualized using a scatterplot and loess curve. The logarithm of the number of days was plotted due to its skewed distribution. The loess curve trend line was flat for the first approximately 300 days, then gradually increased afterwards, suggesting an increasing, non-linear relationship between log days between measurements and change in DFI. Spearman’s correlation coefficient was 0.26 (p < 0.001).

Table 3.

Change in DNA fragmentation index per patient in the different age categories according to time between measurements

Parameter	< 1 year	1–3 years	> 3 years	P valuea
40 years or younger
Number of patients	76	42	21
DFI difference per patient	− 8.4±16.0	−4.7±12.7	4.1±11.0	0.003*
DFI % change per patient	− 16.6±31.0	−2.1±40.9	3.2±8.4	0.004*
40–50 years
Number of patients	42	23	10
DFI difference per patient	− 5.8±12.8	− 5.5±14.7	13.4±13.9	0.002*
DFI % change per patient	− 10.2±31.4	− 9.7±30.3	48.0±59.5	< 0.001*
50 years or older
Number of patients	6	3	2
DFI absolute difference per patient	− 1.1±14.9	6.4±9.3	− 11.6±2.7	0.3
DFI % change per patient	− 0.9±32.4	14.8±20.8	− 26.4±8.6	0.26

Data presented as mean ± SD

DFI DNA fragmentation index

^aKruskal-Wallis test * Statistically significant (p < 0.05)

Fig. 1.

Scatterplot visualizing relationship between days between DNA fragmentation index (DFI) tests and change in DFI (final–baseline DFI) with loess curve. Time between DFI tests was logarithmically transformed

An additional analysis with the 3–5 years and 5+ years groups collapsed was done due to the small number of patients in the latter group (Suppl. Table 1). This analysis demonstrated a difference in baseline DFI—lower in the > 3 years group compared to the 1–3 years and < 1 year (30.4 ± 16.5 vs. 40.4 ± 18.3 and 42.9 ± 16.0). Absolute and relative change in DFI also differed between groups—a decrease in DFI was noted in the < 1 year (− 7.1 ± 14.9 and − 13.7 ± 31.3%, respectively) and 1–3 years (− 4.5 ± 13.4 and − 3.9 ± 37.0%, respectively), while an absolute and relative increase in DFI was noted in the > 3 years (+ 5.9 ± 13.0 and + 25.9 ± 52.3%, respectively) group (p < 0.001 for both). Pairwise comparisons for these three variables showed a significant difference (p < 0.05) for the < 1 year vs. > 3 years and 1–3 years vs. > 3 years for all three parameters.

The relationship between age at baseline DFI and DFI change was tested using the Spearman correlation coefficient and demonstrated no significant correlation between these parameters (R = 0.07, p = 0.27) (Suppl. Figure 1).

Univariate regression models demonstrated that baseline DFI (categorized as ≤ 30% vs. > 30%), time between measurements (categorized as < 1 year, 1–3 years, and > 3 years), medication use, diabetes status, and alcohol consumption were associated with DFI change. The results of the multivariate regression models containing these parameters, excluding diabetes status due to the small number of patients with diabetes (n = 6), are displayed in Table 4. Baseline DFI of ≥ 30% (estimate = 20.34 [95% CI: 15.01, 25.67], p < 0.001) and > 3 years between tests (estimate = 7.00 [95% CI: 0.52, 13.49], p < 0.001) were each associated with larger increases of DFI.

Table 4.

Multivariate linear regression model predicting change in DNA fragmentation index values

Parameter	Estimate	95% CI	P value
Baseline DFI
≤ 30%	Reference
> 30%	20.34	15.01, 25.67	<0.001*
Medication use	1.78	− 2.93, 6.49	0.46
Alcohol consumption	-3.65	− 7.95, 13.49	0.10
Time between DFI tests
<1 year	Reference
1–3 years	1.99	− 2.82, 6.80	0.42
3+ years	7.00	0.52, 13.49	0.03*

DFI DNA fragmentation index. Reference = reference group for listed predictor * Statistically significant (p < 0.05)

Additionally, univariate and multivariate logistic regression models for prediction of categorical increase of 5% from the baseline DFI showed that the time parameter of > 3 years between tests (OR = 4.66 [95% CI: 1.41, 15.38], p = 0.01) is independently predictive of such DFI increase (Suppl. Table 2). Similar models done to identify factors predictive of a 20% increase in DFI demonstrated that only alcohol consumption (OR = 0.40 [95% CI: 0.17, 0.95], p = 0.03) was significant, while the time parameter of > 3 years between tests did not reach statistical significance (p = 0.051) (Suppl. Table 2).

The subgroup of patients with normal levels of DFI < 30% was additionally assessed by time periods (Suppl. Table 3). The basic characteristics were similar between the time groups. Further analysis showed similar increase in DFI over time (p = 0.05).

Discussion

In this study, we evaluated the intra-individual changes in DFI levels over time. This is the first study to address this in the population of infertile men over long time periods, with studies in DFI changes in individuals to potentially develop predictive models for DFI change. Our results demonstrate a significant improvement in DFI in the first 3 years from the first male infertility visit (with less than a third of patients having higher DFI) and a shift towards higher DFI levels after 3 years (with two-thirds of patients having increased DFI). Moreover, we demonstrated a positive correlation of time between DFI tests and intra-personal DFI increase. A regression analysis for prediction of this change supported this finding by showing that the only two parameters predictive of an increase in DFI were baseline DFI and time of 3 years or more between tests.

The regression analysis was unable to identify factors to predict the level of change in DFI in an individual man apart the abovementioned factors. While there are a wide range of changes in the individual’s DFI levels, based upon this study, we were unable to develop a predictive model to identify an individual man's change in DFI values.

Variability in semen parameters (volume, concentration and motility) was previously investigated among men, showing significant variability in each parameter in both infertile men and normal semen donors [32]. Several studies have investigated the intra-personal variability (coefficient of variant) of sperm DNA fragmentation, comparing samples taken shortly apart, among other semen parameters [33–35]. These studies showed that though biological variation does exist among fertile and infertile men, sperm DNA fragmentation or denaturation has lower biological variability than classical semen parameters. A recent study by Punjabi et al. [36] evaluated intrapersonal variability in sperm DFI over time. This prospective longitudinal study, which eventually examined 94 patients with at least two DFI tests, focused on variability in DFI and accordingly the mean time interval between samples was 80 days. It included unexplained infertility and mild men subfertility (sperm count >5 million/ml) cases. The authors concluded that one out of three men fluctuate between the high and the low SDF groups evaluated by TUNEL assay. Our study, however, aimed to estimate trends over longer periods of time in the entire population of infertile men, with significantly longer time between tests (mean of 586 days). The evaluation of the unique group of men with two DFI tests months and years apart provides an interesting insight on the possible physiology of sperm DNA fragmentation formation, as we have shown for the under 3 years groups (DFI decrease) compared to those of more than 3 years (DFI increase).

The common practice in our male infertility clinic is to recommend lifestyle changes and advise to start multivitamins and fertility supplements (that include oral antioxidant). In a previous study conducted in our center [37], a 3-month course of antioxidant therapy was shown to reduce DFI in 49% of patients who had initial high DFI (> 30%), with an improvement of 10.1% between the initial and post-treatment DFI (46.8 ± 14.1% vs. 36.7 ± 16.6%, p < 0.001). Though speculative, this might explain the decrease in DFI in the 3 first years, when this effect was present and patients were more likely to adhere to recommendations, as opposed to longer follow-up of more than 3 years which is more likely to represent the natural history of changes in sperm DFI. One can also hypothesize that an increase in DFI may have started earlier than 3 years but concealed by the positive lifestyle modifications during that time. Nevertheless, investigation of the longer time intervals demonstrated that over longer periods of time, DFI does eventually and gradually increase.

Our analyses showed no correlation between the age at baseline DFI measurement and the change in DFI over time. While age itself involves biologic and physiologic changes that may affect sperm quality and DNA integrity [38, 39], this process is gradual and varies between individuals. Advanced age additionally entails longer duration of exposure to physiologic factors negatively associated with higher DFI such as stress, weight gain, acute and chronic inflammation processes, environmental pollutants and possible unhealthy lifestyle factors such as smoking, alcohol and recreational drugs use [15]. This chronic exposure itself contributes to decreased sperm quality and breakage of sperm DNA, and this process might also be reflected in our findings of correlation between time and increase in DFI, irrelevant of baseline age.

Our study has several limitations, the main being the relatively small sample size and the lack of fertility outcomes in each time point according to DFI. Second, the long study period of more than 15 years did not enable us to account for lifestyle changes, fertility supplements intake and interventions (such as varicocelectomy) done to improve DFI levels during the time between tests and data was lacking regarding the indication for the repeat DFI measurements. Baseline DFI levels were higher in the < 1 year and 1–3 years groups. Patients that only had short follow-up on DFI might represent higher reproductive potential and positive outcomes, refraining them to return to follow-up. Moreover, these time groups demonstrated significantly improved DFI at follow-up and perhaps their high baseline DFI was related to factors easily manageable by improving harmful exposure, lifestyle, intake of multivitamins, and fertility supplements or varicocele repair and therefore resulting in better outcomes after elimination of such factors as part of the clinic recommendations.

The fact that only infertile men were included in this study may hamper the ability to generalize our results and infer conclusions regarding the physiology of DFI alteration over time. However, DFI is rarely done in fertile men, and almost never done more than once in this population; therefore, the infertile men population is the population of interest for such investigation

Our findings add new data on the evolution of SDF over a long period of time. Testing infertile men’s DFI levels at first evaluation may contribute to personalized consult regarding future reproductive outcomes and family planning, especially for men with initial borderline high DFI levels. Moreover, our study is the first to demonstrate the long-term change in DFI, reflecting on sperm quality, in the individual level and not only by age in the population level. These findings advocate further studies aimed to prospectively investigate changes in DFI with time in a well-characterized infertile men and in men with DFI within the normal reference range in order to identify risk factors for DFI increase in individual men over time and correlate it with reproductive outcomes. Moreover, it is important to explore not only semen parameters but also factors such as abstinence period and other factors that might affect DFI levels.

Conclusion

DFI levels gradually increase over time, compared to the baseline levels, and reach an average of 10.8% elevation per patient after 5 years. There is a need to study more well-characterized men with infertility to determine if we could identify factors predictive of DFI changes in individual men over time; however, this information may be useful for future research and clinical practice.

Author contributions

G. K. and M. S. K. contributed substantially to the conception and design of the study, the analysis and interpretation of data, and the drafting of the article. S. L. contributed to the conception and design of the study, the acquisition and interpretation of data, and the drafting. K. C. L. and E. D. G. contributed to the analysis and interpretation of data and the revision of the article. B. A. contributed to the conception and design of the study and interpretation of data. K. L. contributed to the design of the study, analysis, and interpretation of the data. B. M. contributed to the acquisition, analysis, and interpretation of the data. K. J. contributed substantially to the conception and design of the study, the acquisition, analysis, and interpretation of the data, and the drafting and revision of the article. All authors approved the final version of the study.

Data and/or code availability

The datasets and materials used in this study may be available from the corresponding author upon reasonable request.

Declarations

Ethical approval

All participants have signed the IRB-approved informed consent form. The Research Ethics Board of the Mount Sinai Hospital approved data collection (reference number 05-0161-E) and analysis (reference number 07-0032-E). The dates of approval were October 18, 2005, and October 30, 2007, respectively.

Informed consent

The IRB-approved informed consent was provided by all participants prior to filling out the computer-based survey after upon first visit to the clinic.

Competing interests

The authors declare no competing interests.