Effectiveness of Exercise Training on Male Factor Infertility: A Systematic Review and Network Meta-analysis

Abstract

Context:

Mounting evidence from the literature suggests that different types of training interventions can be successful at improving several aspects of male reproductive function in both fertile and infertile populations.

Objective:

The aim of this study was to evaluate the effectiveness of exercise training on male factor infertility and seminal markers of inflammation.

Data Sources:

We searched PubMed, CISCOM, Springer, Elsevier Science, Cochrane Central Register of Controlled Trials, Scopus, PEDro, Ovid (Medline, EMBASE, PsycINFO), Sport Discus, Orbis, CINAHL, Web of Science, ProQuest, and the ClinicalTrials.gov registry for randomized controlled trials (RCTs) that analyzed the impacts of selected types of exercise interventions on markers of male reproductive function and reproductive performance.

Study Selection:

A total of 336 records were identified, of which we included 7 trials reporting on 2641 fertile and infertile men in the systematic review and network meta-analysis.

Level of Evidence:

Level 1 (because this is a systematic review of RCTs).

Data Extraction:

The data included the study design, participant characteristics, inclusion and exclusion, intervention characteristics, outcome measures, and the main results of the study.

Results:

The results of network meta-analysis showed that, compared with a nonintervention control group, the top-ranking interventions for pregnancy rate were for combined aerobic and resistance training (CET) (relative risk [RR] = 27.81), moderate-intensity continuous training (MICT) (RR = 26.67), resistance training (RT) (RR = 12.54), high-intensity continuous training (HICT) (RR = 5.55), and high-intensity interval training (HIIT) (RR = 4.63). While the top-ranking interventions for live birth rate were for MICT (RR = 10.05), RT (RR = 4.92), HIIT (RR = 4.38), CET (RR = 2.20), and HICT (RR = 1.55). Also, with the following order of effectiveness, 5 training strategies were significantly better at improving semen quality parameters (CET > MICT > HICT > RT > HIIT), seminal markers of oxidative stress (CET > MICT > HIIT > HICT > RT), seminal markers of inflammation (CET > MICT > HIIT > RT > HICT), as well as measures of body composition and VO₂max (CET > HICT > MICT > HIIT > RT).

Conclusion:

The review recommends that the intervention with the highest probability of being the best approach out of all available options for improving the male factor infertility was for CET.

Infertility affects an estimated 10% to 15% of couples in industrialized countries. ²¹ According to the World Health Organization (WHO), about 60 to 80 million couples currently suffer from infertility worldwide, ¹⁸ and approximately 50% of all infertility cases are attributed to male-related factors, in particular, poor semen quality, ² called male factor infertility. Male factor infertility is characterized by the existence of suboptimal sperm parameters in the male partner of a couple of childbearing age and is presently defined as the inability to conceive a child with frequent and unprotected sexual intercourse in the fertile phase of the menstrual cycles for a year or longer. ²¹ A wide variety of risk factors such as chronic health problems, injuries, illnesses, and lifestyle factors can deteriorate the male reproductive function. ⁶ In addition, proinflammatory cytokines, which are critical intracellular communicators, are fundamentally implicated in normal reproductive physiology, and local or systemic disturbance of these factors can have a detrimental impact on testicular function and, ultimately, fertility. In this sense, increased levels of proinflammatory cytokines in seminal plasma have been shown to reduce the number of spermatozoa, progressive motility, sperm morphology, and sperm viability and aggravate sperm membrane lipid peroxidation by elevating reactive oxygen species (ROS) production. ¹⁷ It is also widely recognized that oxidative damage to spermatozoa plays a role in the pathogenesis of male factor infertility. Several studies have connected oxidative stress to poor sperm fertilization potential and embryonic development, pregnancy loss, and birth abnormalities. ³ Furthermore, compared with males with a normal body mass index (BMI), men with overweight or obesity have more inferior sperm quality, including reduced sperm concentration, sperm motility, acrosome reaction decline, more significant sperm DNA damage, and decreased embryo implantation rates. As a result, obesity has been linked to a 20% rise in occurrences of subfertility and infertility.^4,22 Considering the aforementioned facts, however, it seems that living a physically active lifestyle can alleviate their repercussions and accordingly bring about a more favorable outcome. ⁸

Benefits of different exercise interventions on markers of male reproductive function have already been investigated in fertile and infertile human participants. There have been randomized controlled trials (RCTs) reporting on the positive effects of different exercise modalities in the form of moderate-intensity continuous training (MICT), ⁹ resistance training (RT), ¹⁰ combined aerobic and resistance training (CET), ⁷ high-intensity continuous training (HICT), ¹⁹ and high-intensity interval training (HIIT) ⁸ on markers of male reproduction. These RCTs concluded that those exercise stimuli are effective in improving a number of indicators of male fecundity, including semen quality parameters as well as pregnancy and live birth rate, all of which reveal positive outcomes for reproductive function and are important from the fertility point of view. According to the authors, these beneficial effects have been attributed to the training-induced improvements in seminal markers of inflammation and antioxidative stress as well as improvements in body composition measures.

To the best of our knowledge, to date, neither systematic literature review nor network meta-analysis has pooled the effects of different training interventions on male reproductive outcomes in both fertile and infertile populations. Obviously, it is critically important to expand a better comprehension of the overall magnitude of the effect, as well as factors associated with exercise training for improving reproductive function and performance in men with infertility problems. Therefore, the primary objectives of this study were to undertake a systematic literature review with network meta-analysis of RCTs to evaluate the effectiveness of 1 or more of the selected types of exercise interventions (MICT, RT, CET, HICT, and HIIT) in the prevention and treatment of male factor infertility as well as to provide clinicians with a ranking of treatments to inform them of the treatment effects of exercise training and physical activity in general in men with infertility problems.

Methods

Search Methods for Identification of Studies

A systematic search of the literature was performed by 2 independent investigators on the electronic databases of PubMed, CISCOM, Springer, Elsevier Science, Cochrane Central Register of Controlled Trials, Scopus, PEDro, Ovid (Medline, EMBASE, PsycINFO), SPORTDiscus, Orbis, CINAHL, Web of Science, ProQuest, and the ClinicalTrials.gov registry from the earliest record in each database to March 30, 2020, restricted to either randomized or controlled trials. The terms used for the search were derived from a combination of the following words: exercise, exercise training, physical activity, sports, exercise therapy, sports therapy, strength, resistance, aerobic, anaerobic, endurance, high intensity, high-intensity interval, combined, concurrent, sperm, spermatozoa, semen, seminal plasma, semen quality, semen parameters, human reproduction, reproductive function, reproductive performance, fertility, sterility, infertility, fecundity, male infertility, male factor infertility, sub-fertility, male sub-fertility, pregnancy, pregnancy rate, reproductive outcomes, live birthrate, and randomized controlled trial. Furthermore, we checked the reference lists from retrieved articles to identify relevant studies, and duplicates were eliminated. Abstracts were checked for eligibility, and full-length reports of potential studies were retrieved for additional eligibility assessment. Disagreements regarding literature search and review were resolved by consensus, and the selection process was entered into a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram (Figure 1). This systematic review was organized, performed, and reported in compliance with the PRISMA ²⁰ and PROSPERO guidelines.

Figure 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram of studies included in the network meta-analysis.

Inclusion Criteria

The inclusion criteria for this systematic review and network meta-analysis were as follows: (1) only RCTs involving healthy adult (age >18 years) human participants and/or infertile patients (with doctor-diagnosed male factor infertility) that examined the effects of supervised and not home-based exercise or physical activity interventions (irrespective of the modality, type, and intensity of the intervention) on markers of male reproductive function and reproductive performance were considered; (2) direct and indirect evidence from RCTs compared 2 or more training modalities and/or an exercise intervention group with a comparative control group (nonintervention, usual care, attention control, and wait-list control); (3) no limitations on language and date of publication; (4) intervention period ≥10 weeks with a follow-up period; and (5) assessment of “primary outcome” markers and indicators—pregnancy rate, live birth rate, semen volume, progressive sperm motility, sperm concentration, sperm morphology, and sperm DNA fragmentation; assessment of “secondary outcome” markers and indicators—seminal markers of inflammation, seminal markers of oxidative stress and antioxidants, body composition measures (weight, BMI, fat%, waist circumference, etc), and cardiorespiratory fitness.

Risk-of-Bias Assessment in Retained Studies

To assess the methodological quality of trials, full copies of included studies were autonomously evaluated by 2 review authors based on the Cochrane Risk-of-Bias Assessment tool for RCTs^13,14 in the following domains: (1) bias arising from the randomization process, (2) bias due to deviations from intended interventions, (3) bias due to missing outcome data, (4) bias in measurement of the outcome, and (5) bias in selection of the reported result (Appendix Figures A1A-A1B, available in the online version of this article). Each domain was classified as being at high, unclear, or low risk of bias. Disagreements on the risk of bias evaluation was settled by discussion between the 2 review authors, and a third investigator if needed.

Data Extraction

For each of the final identified RCTs, the following data were independently extracted by 2 investigators: study characteristics (authors, publication year, country of origin, studied groups, trial design), characteristics of participants and/or patients (sample size, mean age, fertility status, and duration and type of infertility), features of exercise intervention (type, intensity, duration, and frequency of exercise program), and outcome measures. The full text of included RCTs was used for data extraction. The corresponding authors were contacted to get further information in detail, where the provided data in the reports were inadequate to complete the extraction process.

Statistical Analysis

We conducted a network meta-analysis to pool the effect sizes of several modalities obtained from RCTs. We used the R software program Version 3.6.1. to analyze the data and depict the relevant graphs. The netmeta and netmetabin packages were implemented in the software environment. Network graph was depicted to see the overall structure of comparisons. Net split table, which splits our network estimates into the contribution of direct and indirect evidence, enables us to control for inconsistency in specific comparisons in our network. A forest plot was used displaying all comparisons. The total inconsistency and heterogeneity of effect sizes were directly checked using the full design-by-treatment interaction random-effects model. Random-effects model was employed for all data. Standardized mean differences (SMDs) and relative risks (RRs) were reported with their 95% CIs. Publication bias was tested by the Begg-Mazumdar test of the intercept to quantify the bias captured by the funnel plot and to test whether it was statistically significant. Finally, the netrank function implemented in netmeta allows us to generate the rank of treatments from most to least beneficial using P scores. These P scores measure the certainty that one exercise training is better than another one, averaged over all competing exercises.

Results

Characteristics of Included Studies and Participants

Of the 336 citations reviewed, 7 RCTs representing 18 groups (11 exercise, 7 nonintervention control [NON-EX]) and 2641 participants and/or patients (1429 exercise, 1212 NON-EX) met the inclusion criteria in Appendix Table A1 (available online). All the RCTs were conducted in Iran and published in 6 different English-language journals between 2016 and 2020. All the included trials also reported dropouts, estimating sample size before the initiation of their studies, and protocol approval by an ethics committee. Five studies included sedentary married patients with asthenozoospermia, asthenoteratozoospermia, oligoasthenozoospermia, oligospermia, and oligoasthenoteratozoospermia,^7-10,19 and 2 were limited to only sedentary married fertile men.^11,12 None of the studies reported changes in either patients’ dietary intakes or normal daily physical activities and lifestyles during the intervention period other than to comply with the requirements of the study. All the trials reported no history of chronic illness, serious systemic diseases, testicular varicocele, and genital infection; no history of antioxidant use or supplements such as vitamins and medications that could alter the hypothalamic-pituitary-gonadal axis; no history of cigarette and alcohol use in the past 6 months; not working in professions where the activity might influence reproductive capacity; and no relevant previous surgery (eg, vasectomy reversal or varicocele removal). In all the trials, according to the original study authors, as determined by a normal physical examination and routine laboratory tests within the previous year, all the participants were in good health, with regular eating patterns and with no history of depression or illness, and with normal physical and sexual development. All the RCTs reported adequate information with regard to dropouts ranged from 6.1% to 13.8% in the 11 exercise groups (mean ± SD, 8.8 ± 2.5; median, 8.1) and 1.4% to 9.6% in the 7 NON-EX groups (mean ± SD, 5.0 ± 2.9; median 5.4). Reasons for dropping out included exercise adherence, dietary adherence, job change, loss of interest, personal reasons, medication nonadherence, and nonsedentary behavior.

Exercise Program Characteristics

Of the 11 exercise groups, 3 (27.3%) participated in MICT,^9,11,12 2 (18.2%) in HICT,^12,19 2 (18.2%) in HIIT,^8,12 2 (18.2%) in RT,^10,11 and 2 (18.2%) in CET.^7,11 MICT included walking or jogging on a treadmill. HICT and HIIT included treadmill running. RT consisted of exercises to strengthen all major muscle groups, including the upper and lower body. The main exercise program for CET included moderate-intensity aerobic exercise followed by the resistance training protocol. The within-study number of sets, for those trials that involved RT, ranged from 2 to 3, the number of repetitions from 12 to 13; the rest period between sets was 30 seconds for the 3 groups reporting data. The number of RT exercises performed for these 3 groups was 15 exercises. Regarding the delivery of exercise, 11 groups (100.0%) from 7 RCTs participated in supervised exercises (Appendix Table A1, available online).

Risk-of-Bias Assessment

An outline of the risk-of-bias assessment findings is presented in Appendix Figures A1A-A1B (available online). There were 4 (57.1%) RCTs^7,10-12 with low risk of bias and 3 (42.9%)^8,9,19 RCTs with some concerns on random sequence generation. In addition, 14.3% of the studies ¹¹ and 85.7% of the studies,^7-10,12,19 respectively, were either at a low risk of bias or included some concern for deviation from intended interventions and outcome measurements. Furthermore, 57.2% of the studies^7,9-11 reported low risk of bias, while 14.3% of the studies ¹² and 42.9% of the studies,^8,19 respectively, were included either some concern or a high risk of bias for incomplete outcome data. Finally, all the studies were at low risk of bias with respect to the selection of the reported result.

Data Synthesis

A network plot for the included studies and the most frequent comparisons are represented in Appendix Figure A2 (available online).

Primary and Secondary Outcomes

The pooled estimates of network comparisons on total pregnancy and live birth rates expressed as RR with 95% CIs are summarized in Table 1. In detail, compared with the NON-EX group, with the following order of effectiveness, 5 strategies were significantly better at improving the pregnancy rate (CET [RR, 27.81; 95% CI, 13.34-57.97], MICT [RR, 26.67; 95% CI, 11.18-63.63], RT [RR, 12.54; 95% CI, 5.58-28.19], HICT [RR, 5.55; 95% CI, 2.37-13.00], and HIIT [RR, 4.63; 95% CI, 1.94-11.02]) and live birth rate (MICT [RR, 10.05; 95% CI, 0.71-141.35], RT [RR, 4.92; 95% CI, 0.82-29.52], HIIT [RR, 4.38; 95% CI, 0.72-26.64], CET [RR, 2.20; 95% CI, 0.93-5.17], and HICT [RR, 1.54; 95% CI, 0.68-3.52]) (Table 1). As for male infertility subgroups, the pooled estimates of network comparisons on total pregnancy and live birth rate RRs with 95% CIs are summarized in Appendix Tables A2 and A3 (available online).

Table 1.

Relative risk (RR) with 95% CIs for the network comparisons on total pregnancy and live birth rates

Outcomes	No. of Studies	Sample Size	RR	95% CI
Pregnancy rate
CET vs MICT	1	129	1.04	(0.33-3.25)
CET vs NON-EX	2	651	27.81	(13.34-57.97)
CET vs RT	1	126	2.22	(0.74-6.61)
HICT vs HIIT	1	127	1.20	(0.35-4.04)
HICT vs MICT	1	128	0.21	(0.06-0.70)
HICT vs NON-EX	2	520	5.55	(2.37-13.00)
HIIT vs MICT	1	131	0.17	(0.05-0.59)
HIIT vs NON-EX	2	553	4.63	(1.94-11.02)
MICT vs NON-EX	3	650	26.67	(11.18-63.63)
MICT vs RT	1	127	2.12	(0.65-6.97)
RT vs NON-EX	2	534	12.54	(5.58-28.19)
Live birth rate
CET vs MICT	1	129	0.22	(0.01-3.52)
CET vs NON-EX	2	651	2.20	(0.93-5.18)
CET vs RT	1	126	0.45	(0.06-3.26)
HICT vs HIIT	1	127	0.35	(0.05-2.56)
HICT vs MICT	1	128	0.15	(0.01-2.45)
HICT vs NON-EX	2	520	1.55	(0.68-3.52)
HIIT vs MICT	1	131	0.44	(0.02-10.70)
HIIT vs NON-EX	2	553	4.38	(0.72-26.64)
MICT vs NON-EX	3	650	10.05	(0.71-141.35)
MICT vs RT	1	127	2.04	(0.08-49.85)
RT vs NON-EX	2	534	4.92	(0.82-29.52)

CET, combined aerobic and resistance exercise training; HIIT, high-intensity interval training; HICT, high-intensity continuous training protocol; MICT, moderate-intensity continuous training; NON-EX, nonintervention control; RT, resistance training.

Pairwise network comparisons, using both direct and indirect evidence, are presented for semen quality parameters (Appendix Table A4, available online, and Figure 2), seminal markers of inflammation (Appendix Table A5 and Figure A3, available online), and oxidative stress (Appendix Table A6 and Figure A4, available online) as well as measures of body composition and VO₂max (Appendix Table A7 and Figure A5, available online).

Figure 2.

Forest plot showing the results of the network meta-analysis for semen quality parameters. CET, combined aerobic and resistance exercise training; HICT, high-intensity continuous training protocol; HIIT, high-intensity interval training; MICT, moderate-intensity continuous training; NON-EX, nonintervention control; RT, resistance training, SMD, standardized mean difference.

Publication Bias

Begg-Mazumdar test, captured by the funnel plot, was not statistically significant for any of the studied primary and secondary outcomes among the 5 training strategies, except for BMI (P < 0.001), suggesting no substantial publication bias (Appendix Table A8, available online).

Heterogeneity and Consistency Assessment

Significant between-studies heterogeneity was found for changes in percentages of TUNEL positive spermatozoa (I² = 90.7%; Q = 64.66; P = 0.001), superoxide dismutase (I² = 81.7%; Q = 32.70;P = 0.001), total antioxidant capacity (I² = 73.9%; Q = 23.01; P = 0.001), ROS (I² = 82.2%; Q = 33.64; P = 0.001), malondialdehyde (I² = 59.3%; Q = 14.75; P = 0.02), 8-isoprostane (I² = 67.00%; Q = 18.16; P = 0.005), tumor necrosis factor–α (I² = 51.6%; Q = 12.39; P = 0.05), and waist circumference (I² = 78.3%; Q = 27.62; P = 0.001) (Appendix Table A9, available online). Appendix Tables A4 to A7 (available online) also summarize the evidence of consistency within the network at the 5% level using local and global inconsistency tests. The local inconsistency test for assessing inconsistency was also used to identify which treatment comparisons are driving the inconsistency. Node-splitting, first proposed by Dias et al, ⁵ involves separating out the evidence for a particular treatment comparison into the direct and indirect evidence and assessing the discrepancy between them, 1 treatment comparison at a time. The design-by-treatment interaction model of Higgins and Whitehead ¹⁵ did not indicate any global inconsistency in the network.

As shown in Appendix Figures A6 to A12 (available online), the mean age of the participants in each treatment arm (individual points) is close to the overall mean age of the participants in the evidence base (red dotted line). According to the standard deviation (the ± error bars), the variability of ages within each treatment arm appears to be similar as well. Drawing on this analysis, it can be inferred that there is no substantial heterogeneity in age distribution. This analysis was replicated for all possible important effect modifiers found a priori by clinical opinions and a review of previous studies.

To assess the presence of inconsistency, we obtain leverage plots (Appendix Figure A13, available online) where we find that the deviance information criterion (DIC) for the consistency model was marginally smaller than that of the inconsistency model. Overall, we conclude that there was a lack of evidence to suggest inconsistency within the network. Because of excessive zero data, the DIC for primary outcome live birth could not be calculated.

Ranking of Different Training Interventions

The ranking of all selected types of training interventions (CET, MICT, RT, HICT, and HIIT) for each outcome, with NON-EX as the reference group, is presented in Table 2. Briefly, the top-ranking interventions for improving fertility outcomes were for CET, MICT, RT, HICT, and HIIT. More details on the cases are provided in Table 2.

Table 2.

Ranking of different training interventions in the network meta-analysis by P scores for each outcome

	Interventions
Outcomes	CET	MICT	RT	HICT	HIIT	NON-EX
Pregnancy rate	0.89	0.87	0.61	0.34	0.28	0.0001
Live birth rate	0.45	0.82	0.70	0.30	0.66	0.06
Semen volume	0.95	0.84	0.30	0.51	0.07	0.31
Progressive motility	0.97	0.82	0.25	0.51	0.43	0.00
Sperm morphology	0.98	0.76	0.41	0.49	0.34	0.004
Sperm concentration	0.96	0.83	0.45	0.42	0.32	0.004
Number of spermatozoa	0.94	0.83	0.31	0.60	0.19	0.10
TUNEL positive spermatozoa	0.87	0.66	0.51	0.46	0.45	0.07
SOD	0.88	0.76	0.47	0.53	0.34	0.01
Catalase	0.86	0.93	0.38	0.45	0.29	0.07
TAC	0.99	0.80	0.35	0.03	0.50	0.30
ROS	0.81	0.65	0.38	0.63	0.49	0.02
MDA	0.94	0.78	0.26	0.40	0.61	0.001
8-Isoprostane	0.88	0.72	0.42	0.30	0.58	0.09
IL-1β	0.97	0.78	0.49	0.21	0.51	0.05
IL-6	0.95	0.84	0.43	0.39	0.37	0.01
IL-8	0.91	0.74	0.47	0.23	0.58	0.05
TNF-α	0.91	0.73	0.45	0.41	0.49	0.003
Weight	0.47	0.62	0.20	0.99	0.72	0.00
BMI	0.43	0.74	0.22	0.92	0.70	0.00
Fat%	0.99	0.72	0.59	0.39	0.33	0.00
Waist circumference	0.95	0.67	0.26	0.78	0.34	0.00
VO2max	0.99	0.62	0.42	0.73	0.22	0.00

BMI, body mass index; CET, combined aerobic and resistance exercise training; HICT, high-intensity continuous training protocol; HIIT, high-intensity interval training; IL-1β, interleukin 1-beta, IL-6, interleukin 6; IL-8, interleukin 8; MDA, malondialdehyde; MICT, moderate-intensity continuous training; NON-EX, nonintervention control; ROS, reactive oxygen species; RT, resistance training; SOD, superoxide dismutase; TAC, total antioxidant capacity; TNF-α, tumor necrosis factor alpha; TUNEL, terminal deoxynucleotidyl transferase-mediated fluorescein dUTP nick end labeling.

Discussion

A Synopsis of Findings

It is the first network meta-analysis investigating the pooled effects of different exercise interventions on male reproductive function. Having included all studied interventional exercise modalities, our study confirmed the use of exercise interventions as effective prevention and treatment strategies for male factor infertility. In detail, compared with NON-EX group, with the following order of effectiveness, 5 training strategies were significantly better at improving the pregnancy (CET > MICT> RT > HICT > HIIT) and live birth (MICT > RT > HIIT > CET > HICT) rates. In male infertility subgroups, however, the exercise effect on pregnancy and live birth rates were completely different. The exercise effectiveness on pregnancy rate ranked with the following order for asthenozoospermic (MICT > RT > CET > HICT > HIIT), asthenoteratozoospermic (CET > MICT > RT > HIIT> HICT), oligoasthenozoospermic (CET > MICT > RT > HICT > HIIT), oligospermic (CET > MICT > HICT> RT > HIIT), and oligoasthenoteratozoospermic (MICT > CET > HICT > RT > HIIT) subgroups. The exercise effectiveness on the live birth rate ranked with the following order for asthenozoospermic (HICT > HIIT> RT > CET > MICT), asthenoteratozoospermic (RT > HIIT > MICT > CET > HICT), oligoasthenozoospermic (CET > HIIT > RT > HICT > MICT), oligospermic (HICT > RT > HIIT > MICT > CET), and oligoasthenoteratozoospermic (HICT > CET > RT > MICT > HIIT) subgroups. There was insufficient evidence of a difference for the selected types of exercise interventions versus NON-EX group for pregnancy and live birth rates in healthy human participants. Semen quality parameters turned out to be significantly more improved after selected types of exercise interventions as compared with NON-EX group with the following order: CET > MICT > HICT > RT > HIIT. As for secondary outcomes, the exercise effectiveness ranked with the following orders for seminal markers of oxidative stress (CET > MICT > HIIT > HICT > RT), seminal markers of inflammation (CET > MICT > HIIT > RT > HICT), and measures of body composition and VO₂max (CET > HICT > MICT > HIIT > RT). A conservative interpretation of the findings is required because they were based on single studies.

Strengths and Limitations

The present network meta-analysis has several strengths and weaknesses. The major strength of the current study is the application of a network meta-analysis to compare the effectiveness of all available exercise training interventions by combining both direct and indirect evidence to identify the best treatment regimen and to rank all treatments in the evidence network, with the current approach thereby leading to more informed decisions. Furthermore, this network meta-analysis includes seminal markers of inflammation and oxidative stress, as well as measures of body composition and VO₂max, which are the key regulators of the reproductive outcomes. The extensive and overly sensitive search strategy was also applied to retrieve as many appropriate studies as feasible; and, with a high level of agreement, trials were selected independently by 2 reviewers; therefore, the selection bias was less likely. Another major strength is the size of the study population; that is, the included trials had a considerable sample size (range 256-521), which provided the power to evaluate publication bias as well as to detect statistically significant SMDs. Also, in the present study, 5 trials^7-10,19 out of the 7 reported enough evidence on causes of male factor infertility; therefore, based on the current finding we are further able to explore the treatment effects of exercise modalities on reproductive outcomes in different infertility subgroups. The current findings provide invaluable and current information to physicians, researchers, and policy makers regarding the effectiveness of training modalities on reproductive outcomes among at-risk populations and infertile patients.

In this network meta-analysis, we made many efforts to minimize publication bias; yet, in spite of comprehensive retrieval of documents and analysis by 2 independent investigators, this study has some limitations. First, the heterogeneity of some findings was significant, which means there was a discrepancy across the included trials. This could have been triggered by variations in the characteristics of training programs (ie, type, duration, intensity, compliance, etc) across the trials, as well as other features of the included studies such as participants’ characteristics, the small number of relevant trials within the networks for some comparisons, and the baseline levels of studied variables. Two RCTs^11,12 did not report the pregnancy and live birth rates; therefore, because of the lack of direct comparisons and within-design variation, we were unable to check consistency and heterogeneity assumption, respectively, for these outcomes in our network. To achieve this, as an alternative, we utilized the random-effects network meta-analysis model, and the assumptions were checked graphically. This model relaxed the homogeneity among effect sizes. Therefore, the random effects model is used in case of lack of heterogeneity. In addition, there is a concern that this particular study may not relate to already active patients with male factor infertility. However, it should be emphasized that in this network meta-analysis, we rated the training treatments with the highest probability of being the best method in improving male factor infertility out of all available alternatives. Therefore, it suggests that even already very active infertile men and at-risk populations can consider specific exercise modalities and better participate in the top-ranking interventions addressed in the current network meta-analysis, that is, CET, MICT, and RT.

Interpretation and Implications

A network meta-analysis addressing the impact and effectiveness of interventional exercise modalities on pregnancy and live birth rates, as well as semen quality parameters, as the most frequently evaluated and addressed measures of male reproductive function in both the clinical and the nonclinical interventions related to male reproductive research is needed. Also, the inclusion of seminal markers of inflammation and oxidative stress, as well as measures of body composition and VO₂max, as surrogate markers of male reproductive health, might provide one with more precise and broad information with regard to the impact of different exercise modalities on overall male reproductive function. Concerning this, it has been suggested that many cell components, particularly phospholipids in cellular membranes, can be damaged by an imbalance between ROS generation and use. Lipid peroxidation, or oxidative damage to lipids, activates inflammatory signaling pathways, accelerating lipid peroxidation and increasing intracellular oxidative burden. The chain of events includes lipid peroxidation, loss of membrane integrity with enhanced permeability, diminished sperm motility, structural damage to sperm DNA, and apoptosis, with the end consequence being diminished sperm function and fertilization capability. ¹ Previous research has also found a negative relationship between fatness ¹⁶ and indicators of male reproductive function and performance. Furthermore, in both fertile^11,12 and infertile^7-10,19 populations, significant positive correlations were found between maximal oxygen consumption, progressive motility, sperm morphology, sperm concentration, number of spermatozoa, and sperm DNA integrity. Accordingly, the inclusion of such factors in this network meta-analysis adds clinical value to a clinician trying to elucidate fertility questions for men and sheds light on our understanding of how regular exercise training affects human reproduction.

Regarding the conduct of upcoming RCTs, the small number of relevant trials within the networks available for some comparisons suggests that additional study is required to be carried out in the field. Also, it seems that further trials are needed to analyze the dose-response impacts of interventional exercise modalities on male reproductive function and reproductive performance. The results of the current network meta-analysis propose several domains for development in the reporting of RCTs addressing the impacts of interventional exercise studies on male reproductive function. First, according to the Cochrane Risk-of-Bias tool’s outputs, it is proposed that forthcoming RCTs do a better reporting of information on allocation concealment, blinding of outcome assessment, and incomplete outcome reporting. Regarding the blinding of participants and personnel, it needs to be mentioned that, though all trials were at a high risk of bias, it is not practically possible to blind participants in an exercise intervention study. Therefore, this potential risk arises from the nature of the intervention versus any limitations in the study design of an investigative team. In addition, because adverse events, participant preference, and cost-effectiveness are paramount determining factors in the decision of what treatment strategies to suggest for whom, the reporting of such information should also be considered in future trials.

Although further relevant research is required with respect to the impacts of different types of training strategies on male reproduction before any kind of assurance might be established, the findings of the current network meta-analysis suggest that men with infertility could benefit from different exercise interventions to improve fertility outcomes. The top-ranking interventions for primary outcomes were for CET, MICT, RT, HICT, and HIIT. However, the current findings should be interpreted with caution, because all these conclusions are based on either the single small studies or limited direct comparisons. In addition, the significant improvements in the effect sizes for measures of male reproductive function lend support to the promotion of these kinds of interventions either in the clinical or public health settings.

With regard to the current findings, as well as the fact that exercise is possibly a safe activity and the most cost-effective way than other treatment strategies for male factor infertility, it would appear reasonable to propose that infertile men and at-risk populations take part in the top-ranking interventions addressed in the current network meta-analysis, particularly given the myriad other health profits that might be acquired from taking part in such. The findings of the present study are in line with and confirm the recommendations of the WHO ²³ regarding the necessity of keeping a physically active lifestyle to get substantial immediate and long-term health benefits. Taking all the aforementioned facts into account, therefore, for substantial reproductive health benefits, one should consider doing all selected types of exercise interventions (CET, MICT, RT, HICT, and HIIT) to improve fertility outcomes; however, it seems moderate intensity-aerobic exercise alone, strength training alone, and in particular the combination of the 2 would generally be more favorable to lend clinically significant improvements in male factor infertility.

Conclusion

In the setting of couples with male factor infertility, when compared with the NON-EX group, selected types of exercise interventions improved the relative risk of pregnancy rate with the following order: CET > MICT > RT > HICT > HIIT. The top-ranking interventions for live birth rate were for MICT, RT, HIIT, CET, and HICT. In addition, the interventions with the highest probability of being the best approach out of all available options in improving semen quality parameters were for CET, MICT, HICT, RT, and HIIT. This approach should, thus, be considered when formulating clinical recommendations for preventing and treating male factor infertility; but, with respect to the limitations of the present study, the results should be interpreted with caution, given that they were based on single studies.