Abstract
Objective
This systematic review and meta-analysis was conducted to clarify the effects of abstinence duration on semen quality in human studies.
Methods
This review was prospectively registered in PROSPERO and conducted in accordance with PRISMA 2020 guidelines. PubMed (MEDLINE), Scopus, and Web of Science were searched from inception to the most recent update. Human studies evaluating ejaculatory abstinence duration in relation to semen parameters were included. For methodological clarity, studies were classified into (1) single-ejaculate abstinence-comparison studies (paired or independent-group designs) and (2) repeated or frequent ejaculation studies involving consecutive ejaculations. Only single-ejaculate abstinence-comparison studies contributed to the primary meta-analyses. Random-effects meta-analyses were performed for semen volume, sperm concentration, total sperm count, total sperm motility, progressive sperm motility, and sperm morphology. Subgroup and sensitivity analyses were conducted, and risk of bias was assessed using ROBINS-I.
Results
Eighty-two studies were included in the qualitative synthesis, of which 59 contributed to the quantitative meta-analyses. Comparisons involving longer abstinence intervals (e.g., ≥ 4–5 days compared with 2–3 days) were associated with significantly higher semen volume (MD − 1.07 mL, indicating higher volume with longer abstinence) and total sperm count (MD − 80.61 × 10⁶ sperm per ejaculate). In contrast, very short abstinence intervals were associated with significantly higher total sperm motility (MD 4.37%) and progressive sperm motility (MD 3.24%). No consistent overall differences were observed for sperm concentration or morphology across abstinence-interval contrasts. Substantial heterogeneity was observed across outcomes; however, sensitivity analyses generally produced similar directions of association, although substantial heterogeneity remained. The certainty of evidence was low or very low across outcomes, mainly because of substantial heterogeneity, residual confounding, and imprecision.
Conclusions
Across explicit abstinence-interval contrasts, higher-duration intervals were associated with greater semen volume and total sperm count, whereas lower-duration intervals were associated with higher total and progressive sperm motility. The certainty of evidence was low or very low, and these results represent laboratory associations rather than causal effects or universally applicable clinical thresholds. Repeated or frequent ejaculation protocols and reproductive outcomes were synthesized narratively. Abstinence recommendations should therefore be individualized according to the purpose of semen collection and the clinical context.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12958-026-01602-4.
Keywords: Ejaculation, Sexual Abstinence, Semen Analysis, Sperm Motility, Male Infertility
Introduction
Male factors contribute to approximately 40–50% of infertility cases worldwide, and semen analysis remains the cornerstone of male fertility evaluation [1]. Standard semen parameters—semen volume, sperm concentration, total sperm count, motility, and morphology—are widely used to assess male reproductive potential and to guide clinical decision-making in both natural conception and assisted reproductive technologies (ART) [2]. Among the pre-analytical variables influencing semen quality, ejaculatory abstinence duration is one of the most frequently debated yet inconsistently interpreted factors [3].
Current clinical practice and laboratory standards, including those of the World Health Organization (WHO), recommend an abstinence period of 2–7 days before semen collection, primarily to ensure analytical standardization rather than to optimize fertility outcomes [4]. However, this recommendation is largely pragmatic and does not fully account for accumulating evidence suggesting that abstinence duration may exert differential effects on individual semen parameters, potentially influencing reproductive success [5]. Indeed, prolonged abstinence has been associated with increased semen volume and sperm numbers, whereas shorter abstinence intervals have been reported to improve sperm motility and functional competence in certain populations [6].
Over the past several decades, numerous studies have examined the relationship between abstinence duration and semen quality in healthy men, subfertile populations, and men undergoing ART [7]. These studies vary widely in design, population characteristics, abstinence definitions, and laboratory methods, and their findings are often conflicting. While some investigations report improvements in sperm concentration and total sperm count with longer abstinence, others demonstrate reduced motility, increased oxidative stress, or higher sperm DNA fragmentation with prolonged storage [6]. Conversely, short or consecutive ejaculation protocols have been shown to enhance motility and reduce DNA damage in selected clinical settings, despite reductions in semen volume and total sperm output [8].
Importantly, semen quality is a multidimensional construct, and improvements in one parameter may occur at the expense of another [9]. This raises a clinically relevant question: Is there an optimal ejaculatory abstinence duration that balances sperm quantity and functional quality, or should abstinence recommendations be individualized based on clinical context and fertility goals? Despite the large and growing body of literature, no consensus has been reached, and existing reviews are either narrative in nature or limited to specific outcomes, populations, or time frames. In addition to standard semen parameters, evidence on assisted reproductive technology outcomes (including fertilization, pregnancy, and live birth) is considered narratively, as heterogeneous reporting precludes quantitative synthesis.
Importantly, the literature addressing ejaculatory abstinence includes two conceptually distinct experimental frameworks. The first evaluates semen parameters after different single abstinence intervals prior to ejaculation [10]. The second investigates repeated or frequent ejaculation protocols, in which semen parameters are measured across consecutive ejaculates within a short period [11]. Although frequent ejaculation necessarily implies short abstinence intervals, these protocols address a different biological question related to dynamic sperm turnover rather than single abstinence-interval effects. To avoid conflating these distinct research questions, the present review differentiates between single-ejaculate abstinence-comparison studies and repeated-ejaculation studies.
To address this gap, we conducted a systematic review and meta-analysis of human studies to comprehensively evaluate the effects of ejaculatory abstinence duration on key semen parameters. By synthesizing evidence across a wide range of study designs and populations, performing subgroup analyses based on abstinence-interval contrasts, and assessing robustness through sensitivity analyses, this study aims to clarify the direction and magnitude of abstinence-related changes in semen quality. The findings are intended to inform both laboratory practice and clinical counseling, particularly in the context of natural conception and ART. We hypothesized that longer ejaculatory abstinence increases semen volume and sperm quantity, particularly total sperm count, whereas shorter abstinence improves motility-related functional parameters, reflecting a quantity–quality trade-off. Accordingly, the primary objective of this study was to quantify the direction and magnitude of abstinence-related changes in standard semen parameters through meta-analysis, and to summarize advanced sperm function and assisted reproductive technology outcomes narratively where quantitative pooling was not feasible.
Methods
Protocol registration and reporting standards
This systematic review and meta-analysis was prospectively registered in PROSPERO (International Prospective Register of Systematic Reviews); Registration ID: CRD420251268028. The review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement. The protocol prespecified the research question, eligibility criteria, outcomes of interest, and the analytical plan, including random-effects meta-analysis, subgroup analyses by abstinence-interval contrasts, sensitivity analyses, and assessment of publication bias/small-study effects using funnel plots.
Eligibility Criteria
Eligibility criteria were defined a priori using the PICOS framework (Participants, Exposure, Comparators, Outcomes, and Study design). Studies were eligible if they investigated the association between ejaculatory abstinence duration and semen quality parameters in human males. Only original, peer-reviewed human studies were considered for inclusion.
Study design
Eligible studies were classified into three design categories:
within-subject (paired) abstinence-comparison studies, in which semen samples were obtained from the same participant after different abstinence intervals;
independent-group studies, in which different participants were compared across abstinence-duration categories; and.
repeated or frequent ejaculation studies, in which semen parameters were measured across consecutive ejaculations within short time intervals.
Because these designs address different biological questions and differ in susceptibility to confounding, paired studies and independent-group studies were synthesized separately, and repeated-ejaculation studies were excluded from the primary abstinence-duration meta-analyses and summarized narratively.
Classification of evidence streams
To address methodological heterogeneity in the literature on ejaculatory abstinence, eligible studies were categorized into three predefined evidence streams based on study design and exposure structure.
Stream A: Within-subject (paired) abstinence-comparison studies.
These studies obtained semen samples from the same participants after different abstinence intervals. Because each participant serves as his own control, this design minimizes inter-individual confounding related to age, genetics, lifestyle, and baseline semen quality.
Stream B: Independent-group abstinence-comparison studies.
These studies compared semen parameters between different participants categorized by abstinence duration. Such studies are more susceptible to confounding because differences between groups may reflect factors other than abstinence interval.
Stream C: Repeated or frequent ejaculation studies.
These studies examined dynamic changes in semen parameters across consecutive ejaculations (e.g., daily ejaculation protocols or multiple ejaculates within a short period). Although frequent ejaculation necessarily implies short abstinence intervals, these studies address a distinct physiological question related to sequential sperm turnover rather than isolated abstinence-interval effects.
For quantitative synthesis, only Stream A and Stream B studies evaluating single-ejaculate abstinence intervals were included in the primary meta-analyses, and they were analyzed separately by study design. Stream C studies were excluded from the primary meta-analysis and summarized qualitatively, because their exposure structure differs fundamentally from single abstinence-interval comparisons.
Participants
Eligible participants were human males of reproductive age, including healthy or presumed fertile men, normozoospermic men, subfertile or infertile men, and men undergoing fertility evaluation or ART treatment. Studies focusing on specific male-factor infertility phenotypes, such as oligozoospermia, asthenozoospermia, teratozoospermia, oligo-astheno-teratozoospermia (OAT), cryptozoospermia, or men with elevated sperm DNA fragmentation, were included. Studies exclusively involving pediatric populations, non-human subjects, or azoospermic men without reported sperm recovery were excluded.
Exposure: ejaculatory abstinence duration
The primary exposure of interest was single-ejaculate ejaculatory abstinence duration, defined as the time interval between the previous ejaculation and semen sample collection. Studies were eligible if abstinence duration was reported as categorical intervals, analyzed as a continuous variable, or compared between predefined abstinence-duration groups.
For the purposes of quantitative synthesis, only studies evaluating single-ejaculate abstinence-interval comparisons were included in the primary meta-analyses. Studies investigating repeated or consecutive ejaculation protocols (e.g., daily ejaculation or multiple ejaculations within short intervals) were considered conceptually distinct because they assess dynamic changes in semen parameters across serial ejaculates rather than the effect of a single abstinence interval. These studies were therefore retained for qualitative contextualization but were excluded from the primary quantitative meta-analyses.
When studies reported multiple abstinence categories, all relevant contrasts were extracted. Comparisons were harmonized across studies using explicit abstinence-interval contrasts rather than generalized labels such as “short” or “long” abstinence, in order to avoid misclassification of WHO-recommended intervals. Examples of contrasts included ≤ 4 h versus 2–3 days, 2–3 days versus 4–5 days, 2–3 days versus 6–7 days, and 2–3 days versus ≥ 8 days.
For interpretative clarity, abstinence durations of 2–7 days were considered standard-range abstinence, consistent with WHO laboratory recommendations for semen analysis. Durations of ≤ 1 day were considered very short abstinence, whereas ≥ 8 days were considered prolonged abstinence.
Outcomes of interest
Studies were required to report at least one semen quality parameter measured using standardized laboratory methods. Primary outcomes included semen volume, sperm concentration, total sperm count, total sperm motility, progressive sperm motility, and sperm morphology. Secondary outcomes, included for qualitative synthesis where available, comprised total motile sperm count, sperm DNA fragmentation, sperm chromatin maturity or integrity, oxidative stress markers, mitochondrial function, and ART outcomes such as fertilization, implantation, pregnancy, or live birth. ART and reproductive outcomes were not meta-analyzed because of substantial heterogeneity in outcome definitions, denominators (per cycle, per transfer, or per patient), stimulation and laboratory protocols, and limited availability of extractable quantitative data across studies. These outcomes were therefore summarized narratively to contextualize laboratory findings without implying pooled clinical effects. Only studies providing extractable quantitative data, including sample size and measures of central tendency with dispersion, were included in the meta-analyses.
Information sources
A comprehensive literature search was conducted across multiple electronic databases to identify relevant studies evaluating the association between ejaculatory abstinence duration and semen quality in humans. The databases searched included PubMed (MEDLINE), Scopus, and Web of Science (Core Collection), from their inception to the most recent search date. These databases were selected to ensure broad coverage of biomedical, clinical, and reproductive medicine literature.
In addition to database searches, the reference lists of all included studies and relevant review articles were manually screened to identify additional eligible publications that may not have been captured through the electronic search strategy. Only studies published in the English language were considered for inclusion.
Search strategy
A systematic and comprehensive search strategy was developed in consultation with the study objectives to identify all relevant human studies examining the relationship between ejaculatory abstinence duration and semen quality. Controlled vocabulary terms and free-text keywords were combined using Boolean operators to maximize sensitivity while maintaining specificity.
The full electronic search strategies for PubMed (MEDLINE), Scopus, and Web of Science (Core Collection) are presented in Table 1. The search was conducted from database inception through December 1, 2025, and all retrieved records were exported to a reference management software for de-duplication prior to screening.
Table 1.
Electronic database search strategies
| Database | Search strategy |
|---|---|
| PubMed (MEDLINE) | ((“Sexual Abstinence“[MeSH] OR abstinence[tiab] OR “ejaculatory abstinence“[tiab] OR “abstinence period“[tiab] OR “ejaculation frequency“[tiab] OR “ejaculatory frequency“[tiab]) AND (“Semen“[MeSH] OR semen[tiab] OR “semen quality“[tiab] OR sperm[tiab] OR “sperm quality“[tiab] OR “sperm concentration“[tiab] OR “sperm motility“[tiab] OR “sperm morphology“[tiab] OR “total sperm count“[tiab] OR “DNA fragmentation“[tiab]) AND (humans[MeSH])) |
| Scopus | TITLE-ABS-KEY ( (“ejaculatory abstinence” OR “sexual abstinence” OR “abstinence period” OR “ejaculation frequency” OR “ejaculatory frequency”) AND (“semen quality” OR semen OR sperm OR “sperm quality” OR “sperm concentration” OR “sperm motility” OR “sperm morphology” OR “total sperm count” OR “sperm DNA fragmentation”) ) AND (LIMIT-TO (DOCTYPE, “ar”)) AND (LIMIT-TO (SUBJAREA, “MEDI”)) |
| Web of Science (Core Collection) | TS=((“ejaculatory abstinence” OR “sexual abstinence” OR “abstinence period” OR “ejaculation frequency” OR “ejaculatory frequency”) AND (“semen quality” OR semen OR sperm OR “sperm concentration” OR “sperm motility” OR “sperm morphology” OR “total sperm count” OR “DNA fragmentation”)) |
Study selection process
All records retrieved from the electronic database searches were imported into a reference management software, where duplicate citations were identified and removed. The remaining unique records were then screened independently by reviewers based on titles and abstracts to assess potential eligibility according to the predefined inclusion and exclusion criteria.
Studies considered potentially relevant proceeded to full-text review, during which eligibility was evaluated in detail. Articles were excluded at this stage if they did not involve human participants, did not evaluate ejaculatory abstinence duration as an exposure, did not report semen quality outcomes, or lacked sufficient data for extraction. When multiple publications reported overlapping populations, the most comprehensive or methodologically robust study was retained. Any disagreements during the screening or full-text assessment were resolved through discussion and consensus, with consultation of a third reviewer when necessary.
When multiple publications appeared to include overlapping populations, we retained the report with the most complete extractable data for our predefined outcomes and abstinence contrasts. Selection priority was given to studies with the largest sample size, the most detailed abstinence categorization, and the lowest risk of bias. To avoid double-counting, each unique cohort contributed at most one effect estimate per abstinence contrast per outcome in the primary meta-analysis; additional contrasts were included only in prespecified subgroup analyses when they were clearly derived from non-overlapping data.
Data extraction
Data extraction was performed independently by reviewers using a predefined and standardized data extraction form. Extracted information included study identification details (first author, year of publication, country), study design, participant characteristics (sample size, age, fertility status), definitions and categories of ejaculatory abstinence duration, semen analysis methods, and outcomes assessed. For each semen parameter, quantitative data were extracted separately for the lower-duration and higher-duration abstinence intervals within each contrast, including sample size, mean values, and measures of dispersion (standard deviation or standard error, converted to standard deviation when necessary). Where reported, we extracted the World Health Organization (WHO) laboratory manual edition used for semen analysis and the primary analytical approach (manual microscopy versus computer-assisted sperm analysis [CASA]).
When studies reported multiple abstinence categories or repeated-measure designs, all relevant comparisons were extracted and harmonized according to prespecified abstinence-interval groupings for subgroup analyses. If outcome data were incomplete or unclear, the study was excluded from quantitative synthesis for that specific outcome. Any discrepancies in data extraction were resolved by discussion and consensus, with arbitration by a third reviewer when required. Procedures for handling overlapping cohorts and multiple reports, including rules to prevent double-counting of participants, are described in the study selection process.
Because definitions of sperm motility differ across WHO laboratory manual editions, extracted outcomes were classified as total motility or progressive motility whenever possible. When studies reported “motility” without sufficient clarification, the outcome was evaluated individually. Data were pooled only when the motility definition could be reasonably classified; otherwise, the study was excluded from that specific meta-analysis and retained for qualitative synthesis.
Risk of bias assessment
Non-randomized studies were assessed using the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool. Each study was evaluated across seven domains: bias due to confounding, selection of participants, classification of the exposure or intervention, deviations from the intended exposure or intervention, missing data, measurement of outcomes, and selection of the reported result. Each domain was judged as low, moderate, serious, or critical risk of bias, and an overall judgment was assigned according to ROBINS-I guidance. Randomized trials were assessed separately using the Cochrane Risk of Bias 2 tool. Assessments were performed independently by two reviewers, with disagreements resolved through discussion and, when necessary, consultation with a third reviewer.
Data synthesis and statistical analysis
Because the included studies comprised fundamentally different study designs, quantitative syntheses were conducted separately by design type. Within-subject (paired) abstinence-comparison studies and independent-group studies were analyzed independently and were not combined in a single pooled estimate. Studies evaluating repeated or frequent ejaculation protocols were excluded from the primary abstinence-duration meta-analyses and were summarized narratively.
Quantitative synthesis was performed for semen parameters reported by at least two studies with extractable data. Meta-analyses were conducted using random-effects models to account for anticipated clinical and methodological heterogeneity across studies, including differences in study design, participant characteristics, abstinence definitions, and laboratory methods. All statistical analyses were performed using R (meta package). Results are presented as pooled effect estimates with 95% CIs, and statistical significance was defined as a two-sided p value < 0.05.
Effect size measures
For within-subject studies, mean differences and their standard errors were calculated using the reported standard deviation of the within-participant change whenever available. When the within-participant correlation was not reported and could not be derived from available paired summary statistics, a correlation of r = 0.50 was imputed for the primary analysis. Sensitivity analyses were repeated using correlations of r = 0.25 and r = 0.75. Studies lacking sufficient information for paired-effect estimation under these assumptions were excluded from the paired quantitative synthesis. Paired and independent-group studies were synthesized separately.
Heterogeneity assessment
Statistical heterogeneity was assessed using the Cochran Q test and quantified with the I² statistic, representing the percentage of total variability attributable to between-study heterogeneity rather than chance. I² values of approximately 25%, 50%, and 75% were interpreted as low, moderate, and high heterogeneity, respectively. Given the consistently high heterogeneity observed across semen parameters, random-effects models were retained for all primary analyses regardless of heterogeneity magnitude.
Subgroup analyses
Prespecified subgroup analyses were conducted to explore sources of heterogeneity and to examine whether the effect of ejaculatory abstinence differed according to specific abstinence-interval contrasts. Subgroups were defined based on commonly reported abstinence comparisons, including but not limited to ≤ 4 h vs. 2–3 days, ≤ 4 h vs. 4–5 days, ≤ 4 h vs. 6–7 days, > 4 h–1 day vs. longer abstinence, 2–3 days vs. 4–5 days, 2–3 days vs. 6–7 days, and comparisons involving ≥ 8 days. Tests for subgroup differences were performed using a chi-square statistic, with p values < 0.05 indicating statistically significant effect modification. WHO manual edition and laboratory method (manual versus CASA) were considered a priori as potential sources of heterogeneity; however, formal subgroup or meta-regression analyses based on these variables were not feasible because many studies did not consistently report the specific WHO edition or analytical method used.
Sensitivity analyses
Sensitivity analyses were conducted to assess the robustness of pooled estimates. These included leave-one-out analyses, in which each study was sequentially excluded to evaluate its influence on the overall effect size, and analyses excluding studies judged to have a serious risk of bias based on the ROBINS-I assessment. Additional sensitivity analyses were performed by restricting analyses to within-subject or paired-ejaculate designs, excluding comparisons involving identical abstinence durations, and assessing the impact of alternative statistical assumptions (fixed-effect models). Consistency of findings across these analyses was used to support the stability of the primary results.
Motility analyses were also repeated after excluding studies in which “motility” could not be confidently classified as total or progressive motility and, where feasible, after excluding studies that did not report the WHO manual edition.
Motility analyses were repeated after excluding studies in which the reported motility outcome could not be confidently classified as total or progressive motility. These analyses were conducted separately for within-subject and independent-group studies. For within-subject studies, the paired analyses were additionally repeated using imputed within-participant correlations of r = 0.25, r = 0.50, and r = 0.75.
Publication bias
Potential publication bias and small-study effects were assessed visually using funnel plots for each primary outcome when sufficient studies were available. Funnel plots were examined for asymmetry around the pooled effect estimate and for disproportionate clustering of smaller studies with statistically significant results. Given the observational nature of most included studies and the presence of substantial heterogeneity, funnel plot findings were interpreted cautiously and in conjunction with the overall risk-of-bias and sensitivity analyses.
Certainty of evidence assessment (GRADE)
The certainty of evidence for each primary semen parameter outcome was assessed using a GRADE-informed approach, adapted for quantitative synthesis of observational studies. All outcomes were initially rated as low certainty, consistent with GRADE guidance for non-randomized evidence. Certainty ratings were subsequently downgraded based on explicit, predefined decision rules applied uniformly across outcomes. Downgrading domains included: Inconsistency, assessed using the I² statistic from random-effects meta-analyses. I² values ≥ 75% were considered serious inconsistency, and values ≥ 90% very serious inconsistency. Imprecision, defined as 95% CIs crossing the null effect (mean difference = 0). Publication bias, evaluated using Egger’s regression test when ≥ 10 studies were available for an outcome; p values < 0.10 were considered suggestive of small-study effects. Risk of bias, informed by ROBINS-I assessments; however, no additional downgrade was applied in the primary GRADE synthesis because most studies were judged to have low-to-moderate risk of bias. All GRADE assessments were implemented using a reproducible R-based pipeline, integrating random-effects meta-analysis (REML estimator with Hartung–Knapp adjustment) and automated application of downgrade rules.
Results
Study selection
The electronic database searches identified 1,171 records, and an additional 12 records were identified through other sources, yielding 1,183 records in total. After removal of 370 duplicates, 813 records underwent title and abstract screening. Of these, 726 were excluded, and 87 reports were sought for retrieval. Five reports could not be retrieved, leaving 81 reports for full-text eligibility assessment. All 81 studies met the criteria for inclusion in the qualitative synthesis. The study-selection process and the corresponding reasons are presented in the PRISMA 2020 flow diagram (Fig. 1).
Fig. 1.
PRISMA 2020 flow diagram of study identification, screening, eligibility assessment, and inclusion. The diagram distinguishes studies included in the qualitative synthesis from those contributing to the quantitative meta-analyses and provides reasons for non-inclusion in the quantitative synthesis
Characteristics of included studies
A total of 82 studies met the eligibility criteria and were included in the qualitative synthesis (Table S1), of which 59 studies provided sufficient quantitative data for inclusion in one or more meta-analyses. The included studies were published between 1986 and 2025, reflecting nearly four decades of research on the relationship between ejaculatory abstinence duration and semen quality. According to the predefined evidence-stream framework, the included studies comprised three methodological categories. Stream A included within-subject abstinence-comparison studies evaluating semen parameters after different abstinence intervals in the same participants. Stream B included independent-group studies comparing semen parameters across participants with different abstinence durations. Stream C included studies investigating repeated or frequent ejaculation protocols involving consecutive ejaculates over short intervals. Only Stream A and Stream B studies contributed to the primary quantitative meta-analyses, whereas Stream C studies were summarized qualitatively.
The body of evidence comprised a wide range of observational and experimental designs. These included prospective within-subject studies, retrospective and prospective cohort studies, cross-sectional population-based analyses, and ART-based clinical studies conducted in IVF, ICSI, or IUI settings. A small number of randomized or quasi-randomized studies were also identified. Studies were conducted across multiple regions worldwide, including Europe, Asia, the Middle East, Africa, North America, and South America, indicating broad geographic and clinical representation.
Study populations encompassed healthy normozoospermic men, subfertile and infertile men, and men undergoing ART. Several studies specifically focused on defined male-factor infertility phenotypes, such as oligozoospermia, asthenozoospermia, teratozoospermia, oligo-astheno-teratozoospermia (OAT), cryptozoospermia, or men with elevated sperm DNA fragmentation. Sample sizes varied markedly, ranging from small experimental studies with fewer than 10 participants to large registry-based or population studies including several thousand men or semen samples.
Ejaculatory abstinence was evaluated using heterogeneous definitions across studies. Abstinence durations ranged from very short intervals (≤ 1–4 h, often involving consecutive ejaculates) to prolonged abstinence exceeding 7–14 days, with many studies categorizing abstinence according to WHO-recommended intervals (2–7 days) or subdividing exposure into multiple strata. Several investigations also examined repeated ejaculation protocols or daily ejaculation, while others assessed abstinence as a continuous variable.
All included studies assessed at least one standard semen parameter. Commonly reported outcomes included semen volume, sperm concentration, total sperm count, total and progressive motility, and sperm morphology. A substantial subset of studies additionally evaluated advanced sperm function and quality markers, such as total motile sperm count (TMSC), sperm DNA fragmentation, chromatin maturity, oxidative stress markers, mitochondrial function, and sperm kinematic parameters. Semen analysis was predominantly performed according to WHO laboratory manuals (1992, 1999, 2010, 2021 editions), using manual microscopy and/or computer-assisted sperm analysis (CASA) systems, with ancillary techniques including TUNEL, SCSA, SCD, and COMET assays for DNA integrity.
Semen analysis was most commonly reported as being conducted according to WHO laboratory manuals; however, the specific edition (1992, 1999, 2010, or 2021) was inconsistently specified across studies. Several studies explicitly reported using WHO 2010 or later editions, whereas older studies often referenced WHO criteria without edition detail. Both manual microscopy and CASA-based assessments were used, further contributing to methodological variability.
Across study designs and populations, a consistent pattern emerged whereby longer abstinence durations were associated with increased semen volume, sperm concentration, and total sperm count, whereas shorter abstinence intervals—particularly consecutive or very short abstinence—were more frequently associated with improved sperm motility, reduced sperm DNA fragmentation, and enhanced functional quality, especially in subfertile or ART populations. However, the magnitude and direction of these effects varied by fertility status, outcome measured, and clinical context, underscoring the heterogeneity of the available evidence.
Methodological quality and risk of bias of included studies
Risk of bias was assessed for all included non-randomized studies using the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool. Judgments were made across seven domains: bias due to confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of the reported result.
Overall, the majority of studies were judged to have a low to moderate risk of bias across most domains (Fig. 2). Bias due to classification of interventions and bias due to deviations from intended interventions were consistently rated as low risk, reflecting the objective and clearly defined nature of ejaculatory abstinence duration and semen analysis procedures in most studies. Similarly, bias in measurement of outcomes was predominantly low, as semen parameters were assessed using standardized laboratory methods, most commonly following WHO laboratory manuals and/or validated computer-assisted sperm analysis systems. Overall, 23.2% (19/82) of included studies were judged to be at low risk of bias, 69.5% (57/82) at moderate risk, and 7.3% (6/82) at serious risk of bias according to the ROBINS-I tool. Bias due to confounding was the most frequent concern, with 76.8% (63/82) of studies rated as having moderate or serious risk in this domain, largely reflecting limited adjustment for age, baseline semen quality, lifestyle factors, or underlying infertility diagnoses. In contrast, bias in measurement of outcomes was predominantly low, with 86.6% (71/82) of studies rated as low risk, reflecting the use of standardized semen analysis methods and WHO laboratory manuals. Bias due to classification of interventions and deviations from intended interventions was consistently low across all included studies.
Fig. 2.
Risk of bias assessment of included studies using the ROBINS-I tool: (A) traffic-light plot of domain-level judgments for individual studies and (B) summary plot showing the proportion of studies at low, moderate, and serious risk of bias across domains
In contrast, bias due to confounding and bias in selection of participants were the most frequent sources of concern. Many observational and retrospective studies did not fully adjust for potential confounders such as age, lifestyle factors, comorbidities, baseline semen quality, or prior ejaculatory patterns, resulting in a moderate risk of bias in these domains. A small number of studies were judged to have serious risk of bias, primarily due to limited control of confounding or selective inclusion of specific infertility subgroups.
Bias due to missing data was generally low to moderate; viewed mainly in studies with incomplete reporting of dispersion measures or exclusion of samples with extreme values. Bias in selection of the reported result was rated as moderate in several studies, reflecting selective reporting of specific semen parameters or abstinence comparisons without prior protocol registration.
Methodological quality differed according to study design. Within-subject studies reduced confounding from stable inter-individual characteristics because each participant served as his own control; however, incomplete reporting of within-participant correlations, possible order effects, and variation in sample-collection protocols remained important concerns. Independent-group studies were more susceptible to confounding by age, fertility status, baseline semen quality, lifestyle, comorbidities, and treatment setting. Therefore, estimates from independent-group studies should be interpreted more cautiously than estimates from well-conducted paired comparisons. Overall, the predominance of moderate risk-of-bias judgments limits confidence in the precise magnitude of the pooled associations, even where the direction of association appeared consistent.
Meta-analysis of semen parameters
Semen volume
A total of 59 studies contributed data to the quantitative synthesis of semen volume, encompassing 5,311 samples in the shorter abstinence groups and 3,652 samples in the longer abstinence groups. Random-effects meta-analysis demonstrated that very short abstinence intervals were associated with significantly lower semen volume. Variation in WHO manual editions and laboratory methods across studies likely contributed to the observed between-study heterogeneity.
Overall, the pooled analysis showed a significant mean difference (MD) of − 1.07 mL (95% CI: −1.22 to − 0.93), indicating higher semen volume with longer abstinence intervals (Fig. 3). Considerable between-study heterogeneity was observed (I² = 85.0%, p < 0.0001), justifying the use of a random-effects model. Given this substantial heterogeneity, the pooled mean difference should be interpreted as an average effect across highly diverse study settings rather than a precise estimate applicable to all clinical contexts. Accordingly, the abstinence-interval–specific subgroup analyses provide the most clinically informative insights and are emphasized for interpretation.
Fig. 3.
Forest plot of the meta-analysis of semen volume across explicit ejaculatory abstinence-interval contrasts in single-ejaculate studies, with subgroup analyses according to specific abstinence-interval comparisons. Squares represent study-specific mean differences, with the area of each square proportional to the inverse-variance weight assigned to the study. Horizontal lines represent 95% confidence intervals. Diamonds represent pooled random-effects estimates for the corresponding subgroup; the center of each diamond represents the pooled mean difference and its width represents the 95% confidence interval. The solid vertical line at a mean difference of zero indicates no difference between abstinence intervals. Mean differences were calculated as the outcome after the lower-duration abstinence interval minus the outcome after the higher-duration interval; therefore, negative values indicate a lower outcome value after the lower-duration interval, whereas positive values indicate a higher outcome value after the lower-duration interval
Subgroup analyses based on abstinence-duration comparisons revealed a consistent reduction in semen volume with shorter abstinence across most strata. Comparisons of ≤ 4 h versus 2–3 days, ≤ 4 h versus 4–5 days, and ≤ 4 h versus 6–7 days all demonstrated statistically significant decreases in semen volume with very short abstinence, with pooled MDs ranging from approximately − 1.19 to − 1.25 mL. Similarly, comparisons of > 4 h–1 day versus longer abstinence intervals (4–5 days or ≥ 8 days) showed significantly lower semen volume in the shorter abstinence groups.
When comparing 2–3 days versus 4–5 days and 2–3 days versus 6–7 days, shorter abstinence remained associated with reduced semen volume, although the magnitude of effect was smaller and heterogeneity varied across subgroups. Notably, the test for subgroup differences was not statistically significant (p = 0.276), indicating that the reduction in semen volume with very short abstinence intervals was broadly consistent across different abstinence-duration contrasts.
Sperm concentration
A total of multiple studies comprising 1,284 samples in the shorter abstinence groups and 1,286 samples in the longer abstinence groups were included in the meta-analysis of sperm concentration. Using a random-effects model, no statistically significant overall difference in sperm concentration was observed across abstinence-interval contrasts (MD = − 2.36 × 10⁶/mL; 95% CI: −7.10 to 2.38) (Fig. 4). Substantial heterogeneity was detected across studies (I² = 98.3%, p < 0.0001), indicating marked variability in effect estimates according to abstinence duration contrasts and study characteristics. Given this substantial heterogeneity, the pooled mean difference should be interpreted as an average effect across highly diverse study settings rather than a precise estimate applicable to all clinical contexts. Accordingly, the abstinence-interval–specific subgroup analyses provide the most clinically informative insights and are emphasized for interpretation.
Fig. 4.
Forest plot of the meta-analysis of sperm concentration across explicit ejaculatory abstinence-interval contrasts in single-ejaculate studies, with subgroup analyses according to specific abstinence-interval comparisons. Squares represent study-specific mean differences, with the area of each square proportional to the inverse-variance weight assigned to the study. Horizontal lines represent 95% confidence intervals. Diamonds represent pooled random-effects estimates for the corresponding subgroup; the center of each diamond represents the pooled mean difference and its width represents the 95% confidence interval. The solid vertical line at a mean difference of zero indicates no difference between abstinence intervals. Mean differences were calculated as the outcome after the lower-duration abstinence interval minus the outcome after the higher-duration interval; therefore, negative values indicate a lower outcome value after the lower-duration interval, whereas positive values indicate a higher outcome value after the lower-duration interval
Subgroup analyses revealed considerable variability in the direction and magnitude of effects. In comparisons of ≤ 4 h versus 2–3 days, the pooled effect was not statistically significant, with wide CIs and moderate to high heterogeneity. For ≤ 4 h versus 4–5 days, the pooled estimate suggested a lower sperm concentration in contrasts involving very short abstinence intervals, although the effect did not consistently reach statistical significance across studies and was accompanied by substantial heterogeneity. In the ≤ 4 h versus 6–7 days subgroup, individual studies showed both increases and decreases in sperm concentration with shorter abstinence, resulting in a non-significant pooled effect and very high heterogeneity. Comparisons of > 4 h–1 day versus longer abstinence durations (2–3 days or 4–5 days) tended to show lower sperm concentration in contrasts involving shorter abstinence intervals (≤ 1 d vs. 2–3 d), with some subgroup-level effects reaching statistical significance. In contrasts involving ≥ 8 days of abstinence, estimates were imprecise and based on limited data, yielding wide CIs.
Importantly, the test for subgroup differences was statistically significant (p < 0.0001), indicating that the effect of ejaculatory abstinence duration on sperm concentration differs according to the specific abstinence intervals compared.
Total sperm count
The meta-analysis of total sperm count (TSC) included data from studies comparing shorter versus longer ejaculatory abstinence durations, comprising 2,893 ejaculates in the shorter abstinence groups and 3,591 ejaculates in the longer abstinence groups. Using a random-effects model, shorter abstinence-interval contrasts (≤ 1 d vs. 2–3 d) were associated with significantly lower total sperm count compared with longer abstinence intervals.
Overall, the pooled analysis demonstrated a mean difference (MD) of − 80.61 × 10⁶ sperm per ejaculate (95% CI: −142.12 to − 19.10), indicating a clear advantage of longer abstinence for total sperm output (Fig. 5). Substantial heterogeneity was observed (I² = 96.1%, p < 0.0001), supporting the use of a random-effects approach. Given this substantial heterogeneity, the pooled mean difference should be interpreted as an average effect across highly diverse study settings rather than a precise estimate applicable to all clinical contexts. Accordingly, the abstinence-interval–specific subgroup analyses provide the most clinically informative insights and are emphasized for interpretation.
Fig. 5.
Forest plot of the meta-analysis of total sperm count across explicit ejaculatory abstinence-interval contrasts in single-ejaculate studies, with subgroup analyses by specific abstinence-interval comparisons. Squares represent study-specific mean differences, with the area of each square proportional to the inverse-variance weight assigned to the study. Horizontal lines represent 95% confidence intervals. Diamonds represent pooled random-effects estimates for the corresponding subgroup; the center of each diamond represents the pooled mean difference and its width represents the 95% confidence interval. The solid vertical line at a mean difference of zero indicates no difference between abstinence intervals. Mean differences were calculated as the outcome after the lower-duration abstinence interval minus the outcome after the higher-duration interval; therefore, negative values indicate a lower outcome value after the lower-duration interval, whereas positive values indicate a higher outcome value after the lower-duration interval
Subgroup analyses according to abstinence-duration contrasts revealed important patterns. Comparisons of > 4 h–1 day versus 4–5 days showed a significant reduction in total sperm count with shorter abstinence (MD ≈ − 68.93 × 10⁶), with high heterogeneity. Comparisons of > 4 h–1 day versus ≥ 8 days demonstrated an even larger and statistically significant decrease in total sperm count with shorter abstinence (MD ≈ − 189.58 × 10⁶), indicating a strong cumulative effect of prolonged abstinence on total sperm output. In contrasts of 2–3 days versus 6–7 days, shorter abstinence was again associated with a markedly lower total sperm count, although between-study heterogeneity was very high. Comparisons involving similar abstinence durations showed inconsistent effects and contributed minimally to the overall estimate.
The test for subgroup differences was statistically significant (p = 0.0001), indicating that the magnitude of the effect on total sperm count varies according to the specific abstinence intervals compared.
Overall, these findings demonstrate that longer ejaculatory abstinence consistently results in higher total sperm counts, whereas very short abstinence intervals (≤ 4 h or ≤ 1 d) lead to a substantial reduction in total sperm output, despite variability across study designs and populations.
Sperm motility (Progressive and Total)
The meta-analysis of total sperm motility included 8,742 samples in the shorter abstinence groups and 5,009 samples in the longer abstinence groups. Using a random-effects model, very short abstinence intervals (≤ 4 h or ≤ 1 d compared with 2–3 d) were associated with a statistically significant increase in total motility compared with longer abstinence durations (MD = 4.37%, 95% CI: 2.61 to 6.13) (Fig. 6).
Fig. 6.
Forest plot of the meta-analysis of total sperm motility across explicit ejaculatory abstinence-interval contrasts in single-ejaculate studies, with subgroup analyses by specific abstinence-interval comparisons. Squares represent study-specific mean differences, with the area of each square proportional to the inverse-variance weight assigned to the study. Horizontal lines represent 95% confidence intervals. Diamonds represent pooled random-effects estimates for the corresponding subgroup; the center of each diamond represents the pooled mean difference and its width represents the 95% confidence interval. The solid vertical line at a mean difference of zero indicates no difference between abstinence intervals. Mean differences were calculated as the outcome after the lower-duration abstinence interval minus the outcome after the higher-duration interval; therefore, negative values indicate a lower outcome value after the lower-duration interval, whereas positive values indicate a higher outcome value after the lower-duration interval
Between-study heterogeneity was substantial (I² = 86.3%, p < 0.0001). Given this substantial heterogeneity, the pooled mean difference should be interpreted as an average effect across highly diverse study settings rather than a precise estimate applicable to all clinical contexts. Accordingly, the abstinence-interval–specific subgroup analyses provide the most clinically informative insights and are emphasized for interpretation. Subgroup analyses demonstrated that the motility advantage of very short abstinence intervals was most evident when very short intervals (≤ 4 h) were compared with 6–7 days, and when 2–3 days were compared with ≥ 8 days, whereas comparisons involving intermediate abstinence durations showed smaller or non-significant effects. The test for subgroup differences was statistically significant (p < 0.0001), indicating that the magnitude of the effect varies according to the abstinence intervals contrasted.
For progressive sperm motility, the pooled analysis comprised 9,017 samples in the shorter abstinence groups and 4,502 samples in the longer abstinence groups (Fig. 7). The random-effects model showed a significant improvement in progressive motility with very short abstinence intervals (MD = 3.24%, 95% CI: 1.79 to 4.69).
Fig. 7.
Forest plot of the meta-analysis of progressive sperm motility according to ejaculatory abstinence-interval contrasts in single-ejaculate studies, with subgroup analyses based on specific interval comparisons. Squares represent study-specific mean differences, with the area of each square proportional to the inverse-variance weight assigned to the study. Horizontal lines represent 95% confidence intervals. Diamonds represent pooled random-effects estimates for the corresponding subgroup; the center of each diamond represents the pooled mean difference and its width represents the 95% confidence interval. The solid vertical line at a mean difference of zero indicates no difference between abstinence intervals. Mean differences were calculated as the outcome after the lower-duration abstinence interval minus the outcome after the higher-duration interval; therefore, negative values indicate a lower outcome value after the lower-duration interval, whereas positive values indicate a higher outcome value after the lower-duration interval
Heterogeneity was again high (I² = 86.8%, p < 0.0001). Given this substantial heterogeneity, the pooled mean difference should be interpreted as an average effect across highly diverse study settings rather than a precise estimate applicable to all clinical contexts. Accordingly, the abstinence-interval–specific subgroup analyses provide the most clinically informative insights and are emphasized for interpretation. Subgroup analyses revealed a consistent pattern favoring shorter abstinence, particularly for comparisons of ≤ 4 h versus 6–7 days and 2–3 days versus ≥ 8 days, where pooled effects were positive and statistically significant. Comparisons between similar or adjacent abstinence durations showed attenuated or non-significant effects. As with total motility, the test for subgroup differences was significant (p < 0.0001), confirming effect modification by abstinence-interval contrast.
Sperm morphology
The meta-analysis of sperm morphology included data from 5,081 ejaculates in the shorter abstinence groups and 3,671 ejaculates in the longer abstinence groups. Using a random-effects model, no statistically significant overall difference in the percentage of morphologically normal sperm was observed across abstinence-interval contrasts (MD = 0.43%, 95% CI: −0.07 to 0.93) (Fig. 8).
Fig. 8.
Forest plot of the meta-analysis of sperm morphology according to ejaculatory abstinence-interval contrasts in single-ejaculate studies, with subgroup analyses based on specific abstinence-interval comparisons. Squares represent study-specific mean differences, with the area of each square proportional to the inverse-variance weight assigned to the study. Horizontal lines represent 95% confidence intervals. Diamonds represent pooled random-effects estimates for the corresponding subgroup; the center of each diamond represents the pooled mean difference and its width represents the 95% confidence interval. The solid vertical line at a mean difference of zero indicates no difference between abstinence intervals. Mean differences were calculated as the outcome after the lower-duration abstinence interval minus the outcome after the higher-duration interval; therefore, negative values indicate a lower outcome value after the lower-duration interval, whereas positive values indicate a higher outcome value after the lower-duration interval
Between-study heterogeneity was substantial (I² = 87.2%, p < 0.0001), indicating marked variability across study designs, abstinence comparisons, and populations. Given this substantial heterogeneity, the pooled mean difference should be interpreted as an average effect across highly diverse study settings rather than a precise estimate applicable to all clinical contexts. Accordingly, the abstinence-interval–specific subgroup analyses provide the most clinically informative insights and are emphasized for interpretation.
Subgroup analyses according to abstinence-duration contrasts showed largely neutral and inconsistent effects. Comparisons of ≤ 4 h versus 2–3 days and ≤ 4 h versus 4–5 days showed no significant differences in sperm morphology, with pooled estimates close to zero and minimal heterogeneity. In the ≤ 4 h versus 6–7 days subgroup, a small but statistically significant improvement in morphology was observed with shorter abstinence (MD ≈ 1.19%), although heterogeneity was high. Comparisons of > 4 h–1 day versus 2–3 days, 4–5 days, or ≥ 8 days showed no consistent or statistically significant differences in morphology. Analyses comparing 2–3 days versus longer abstinence durations (6–7 days or ≥ 8 days) similarly demonstrated no clear advantage of either shorter or longer abstinence. Comparisons involving identical abstinence durations confirmed the absence of systematic differences, as expected.
The test for subgroup differences reached statistical significance (p ≈ 0.025), suggesting some variation across specific abstinence contrasts; however, the magnitude of effects was small and clinically modest across all subgroups.
Overall, these findings indicate that, unlike semen volume, total sperm count, and sperm motility, sperm morphology is largely unaffected by ejaculatory abstinence duration. Any observed differences across abstinence intervals were small, inconsistent, and highly heterogeneous, suggesting that morphology is a relatively stable semen parameter with respect to abstinence length.
Advanced sperm function outcomes
In addition to standard semen parameters, several included studies evaluated advanced sperm function and quality markers, including sperm DNA fragmentation, oxidative stress indices, chromatin maturity or integrity, mitochondrial function, and selected kinematic parameters. These outcomes were not suitable for quantitative meta-analysis because of heterogeneous assay methods, reporting formats, and abstinence-interval definitions. A structured overview of advanced sperm function outcomes and assay types is provided in Table S2.
Sperm DNA fragmentation
Multiple studies employing assays such as sperm chromatin structure assay (SCSA), TUNEL, sperm chromatin dispersion (SCD), or COMET reported lower DNA fragmentation indices following shorter ejaculatory abstinence or consecutive ejaculation protocols, particularly in men with elevated baseline DNA damage or subfertile phenotypes. However, the magnitude and consistency of these effects varied across assays and study designs.
Oxidative stress and mitochondrial function
A smaller number of studies assessed oxidative stress markers and mitochondrial activity, reporting reduced oxidative stress and improved mitochondrial membrane potential after shorter abstinence intervals. These findings are biologically plausible given reduced epididymal storage time but remain limited by small sample sizes and heterogeneous methodologies.
Chromatin maturity and related markers
Studies evaluating chromatin maturity or protamination generally reported neutral or modest improvements with shorter abstinence, although results were inconsistent and assay-dependent.
Repeated or frequent ejaculation studies
Several included studies investigated repeated or frequent ejaculation protocols rather than isolated abstinence-interval comparisons. These studies typically involved daily ejaculation or multiple ejaculations within the same day and assessed dynamic changes in semen parameters across consecutive ejaculates. Because such protocols address a different physiological question from single abstinence-interval comparisons, these studies were excluded from the primary meta-analyses and are summarized narratively. Most reported progressive reductions in semen volume and total sperm count across consecutive ejaculations, accompanied by improvements in motility or reductions in DNA fragmentation in some populations.
Reproductive and ART outcomes
A subset of included studies reported reproductive outcomes in assisted reproductive technology (ART) settings, including fertilization rate, embryo development, implantation, clinical pregnancy, and live birth. However, these outcomes were reported using heterogeneous denominators (per cycle, per oocyte, per embryo transfer, or per patient), variable abstinence contrasts, and differing ART protocols. Consequently, quantitative meta-analysis of reproductive outcomes was not feasible.
Narratively, some studies suggested that shorter abstinence or second-ejaculate strategies may improve fertilization or embryo development in selected populations, particularly in men with elevated sperm DNA fragmentation. However, evidence regarding clinical pregnancy or live birth remained inconsistent and limited.
Sensitivity analyses
After exclusion of studies with ambiguous motility definitions, lower-duration abstinence intervals remained associated with higher total sperm motility in paired studies (Stream A: MD = 5.54% points; 95% CI 1.78 to 9.29; I² = 73.1%) and in independent-group studies (Stream B: MD = 3.49% points; 95% CI 0.46 to 6.52; I² = 90.7%). For progressive sperm motility, the association remained positive but was not statistically significant in paired studies (Stream A: MD = 2.30% points; 95% CI − 0.94 to 5.54; I² = 70.5%) and remained statistically significant in independent-group studies (Stream B: MD = 3.13% points; 95% CI 1.25 to 5.00; I² = 80.4%). These sensitivity analyses indicated that the principal directions of association were maintained overall; however, the paired-study estimate for progressive sperm motility was imprecise and its confidence interval crossed the null.
Publication bias
Publication bias and small-study effects were assessed visually using funnel plots for each primary semen parameter included in the meta-analysis (Fig. 9). Visual inspection focused on the symmetry of study effects around the pooled mean difference and the distribution of smaller studies across significance contours.
Fig. 9.
Funnel plots assessing publication bias and small-study effects for semen parameters: (A) semen volume, (B) sperm concentration, (C) total sperm count, (D) total sperm motility, (E) progressive sperm motility, and (F) sperm morphology
For semen volume (Fig. 9A), the funnel plot showed mild asymmetry, with a relative scarcity of small studies reporting effects close to the null. This pattern is consistent with the substantial heterogeneity observed in the meta-analysis and may reflect true between-study variability rather than selective publication.
For sperm concentration (Fig. 9B) and total sperm count (Fig. 9C), the funnel plots demonstrated noticeable asymmetry, particularly among smaller studies, suggesting the presence of small-study effects. These patterns may be attributable to differences in study design, population characteristics, abstinence categorization, or selective reporting of statistically significant findings in smaller cohorts.
In contrast, funnel plots for total sperm motility (Fig. 9D) and progressive sperm motility (Fig. 9E) appeared largely symmetrical, with most studies distributed evenly around the pooled effect estimate and within the expected confidence limits. This visual symmetry suggests a lower likelihood of publication bias for motility outcomes, despite the high heterogeneity observed.
For sperm morphology (Fig. 9F), the funnel plot was also approximately symmetrical, with effect estimates clustered around the null value, consistent with the absence of a significant pooled effect and supporting a low risk of publication bias for this outcome.
Certainty of evidence across semen parameters
The certainty of evidence for the association between ejaculatory abstinence duration and semen quality outcomes was assessed. For semen volume, evidence from 26 studies (9,145 samples) demonstrated a consistent reduction with shorter abstinence (MD − 1.12 mL; 95% CI − 1.29 to − 0.96). Despite a precise estimate, substantial heterogeneity was observed (I² = 85.6%), leading to downgrading for inconsistency. The overall certainty of evidence for semen volume was rated as low. For sperm concentration, data from 25 studies (2,662 samples) showed no statistically significant difference between shorter and longer abstinence durations (MD − 2.03 × 10⁶/mL; 95% CI − 6.87 to 2.81). The evidence was downgraded due to very high heterogeneity (I² = 98.2%) and imprecision, resulting in very low certainty. For total sperm count, 27 studies (12,228 samples) indicated significantly lower total sperm counts with shorter abstinence (MD − 85.2 × 10⁶ per ejaculate; 95% CI − 131 to − 39). Certainty was downgraded because of very serious inconsistency (I² = 97.2%) and evidence of small-study effects (Egger p = 0.0007), yielding very low certainty. For total sperm motility, evidence from 39 studies (13,809 samples) showed higher motility with shorter abstinence (MD + 4.53%; 95% CI 2.81 to 6.25). Although heterogeneity was substantial (I² = 85.8%), the effect estimate was precise and publication bias was not clearly detected. The certainty of evidence was therefore rated as low. For progressive sperm motility, 31 studies (13,519 samples) demonstrated a consistent improvement with shorter abstinence (MD + 3.24%; 95% CI 1.79 to 4.69). Despite high heterogeneity (I² = 86.8%), the direction and magnitude of effect were stable, resulting in low-certainty evidence. For sperm morphology, 33 studies (8,810 samples) showed no significant difference between abstinence durations (MD + 0.43%; 95% CI − 0.07 to 0.92). The evidence was downgraded for serious inconsistency (I² = 86.4%) and imprecision, leading to very low certainty.
Overall, the certainty of evidence was low or very low across the primary outcomes. The principal reasons for reduced certainty were substantial or very serious inconsistency, residual confounding in independent-group studies, incomplete reporting of within-subject variance, imprecision for sperm concentration and morphology, and possible small-study effects for total sperm count. Consequently, the pooled estimates should be interpreted as average directional associations across heterogeneous populations and laboratory protocols rather than as precise or universally applicable clinical thresholds.
Discussion
This systematic review and meta-analysis provides a comprehensive updated synthesis of human evidence of human evidence on the impact of ejaculatory abstinence duration on semen quality. By integrating data from 82 studies spanning nearly four decades and pooling quantitative results from 59 studies, we demonstrate a clear and biologically coherent pattern: longer abstinence durations increase semen volume and total sperm count, whereas shorter abstinence intervals—particularly very short or consecutive ejaculation protocols—consistently improve sperm motility, both total and progressive. In contrast, sperm concentration and morphology showed no consistent overall differences, and their associations with abstinence duration were highly heterogeneous. Importantly, the literature on ejaculatory abstinence encompasses multiple experimental frameworks. In the present study, we distinguished between three evidence streams: [1] single-ejaculate abstinence-comparison studies with within-subject designs [2], single-ejaculate abstinence-comparison studies with independent-group designs, and [3] repeated or frequent ejaculation studies involving consecutive ejaculates. Only the first two categories were included in the primary meta-analyses, whereas repeated-ejaculation studies were interpreted qualitatively because they address a different physiological question.
Importantly, these findings highlight a quantity–quality trade-off in semen characteristics. While prolonged abstinence favors sperm quantity, shorter abstinence appears to enhance functional semen parameters that may be associated with fertilization processes but are imperfect and indirect indicators of reproductive success, especially in subfertile and ART populations. Sensitivity analyses confirmed the robustness of these results across study designs, risk-of-bias strata, and analytical assumptions.
The observed increase in semen volume and total sperm count with longer abstinence likely reflects continued sperm accumulation within the epididymis and accessory sex glands, leading to greater ejaculate volume and sperm output [12]. However, prolonged epididymal storage may also expose spermatozoa to oxidative stress, accumulation of reactive oxygen species, and age-related cellular damage, which can adversely affect motility and DNA integrity [13].
Conversely, shorter abstinence and consecutive ejaculation protocols reduce the storage time of spermatozoa, resulting in a population of younger, metabolically active sperm with superior motility and, in many studies, reduced DNA fragmentation [8, 14]. This mechanism is supported by the consistent improvement in motility and DNA integrity observed after very short abstinence (≤ 4 h or 1 day), particularly in men with oligo-astheno-teratozoospermia or elevated baseline sperm DNA fragmentation [15]. The relative stability of sperm morphology across abstinence intervals suggests that morphological features are largely determined during spermatogenesis and are less influenced by post-testicular storage time [16].
Current WHO guidelines recommend an abstinence period of 2–7 days for semen analysis, primarily to standardize laboratory assessment rather than to optimize fertility outcomes [17]. Our findings partially support this recommendation by showing that intermediate abstinence durations offer a balance between sperm quantity and quality. However, the present meta-analysis extends prior narrative and single-study observations by demonstrating that shorter abstinence may be advantageous for motility-related outcomes, particularly in clinical contexts where sperm function is prioritized over total sperm number.
These results help reconcile previously conflicting findings in the literature and provide quantitative evidence explaining why shorter abstinence has been associated with improved fertilization, implantation, and pregnancy outcomes in some ART studies, despite reductions in semen volume or total sperm count. Importantly, the conclusions of this meta-analysis are based on laboratory semen parameters rather than direct reproductive outcomes, as fertilization, pregnancy, and live birth data were not suitable for quantitative pooling.
Clinical implications
For couples attempting natural conception, prolonged abstinence may increase ejaculate volume and sperm count but may not translate into improved fecundity if motility or DNA integrity is compromised. Our findings suggest that moderate or shorter abstinence intervals (1–3 days) may be sufficient—and potentially preferable—for optimizing sperm functional quality during the fertile window. Importantly, improvements in sperm motility should not be interpreted as direct evidence of increased fertility, as semen parameters are imperfect surrogates and do not reliably predict fertilization, pregnancy, or live birth outcomes at the individual level.
In ART settings, particularly IVF and ICSI, where motility, DNA integrity, and functional competence are critical, shorter abstinence or consecutive ejaculation strategies may be beneficial, especially in men with poor baseline semen quality or elevated sperm DNA fragmentation. These findings support individualized abstinence recommendations rather than a universal abstinence duration for all patients.
Practical recommendations for sperm donors, fertility clinics, and research protocols
The practical interpretation of ejaculatory abstinence depends on the purpose of semen collection. For routine diagnostic semen analysis and sperm-donor eligibility assessment, sample collection after a standardized interval within the WHO-recommended 2–7-day range remains appropriate because the principal objective is comparability with laboratory reference procedures. When repeated samples are collected from the same individual, the same or a narrowly similar abstinence interval should be maintained whenever feasible, and the exact duration should be documented.
In fertility clinics, the collection strategy may be individualized according to the limiting semen characteristic and treatment objective. An interval within the standard 2–7-day range may be preferable when ejaculate volume or total sperm number is the principal concern. Conversely, a very short interval or a second ejaculate after several hours may be considered selectively when motility or sperm DNA integrity is prioritized, particularly in men with impaired baseline semen quality or elevated DNA fragmentation. However, evidence supporting improvements in pregnancy or live birth remains limited, and these strategies should be implemented only within clinician-directed and laboratory-validated protocols.
For sperm-donor programs, prolonged abstinence is not routinely necessary. Donor centers should prioritize a consistent collection schedule, document the exact interval before every donation, and ensure that each specimen meets center-specific thresholds for sperm concentration, motility, infectious-disease screening, and post-thaw performance. Collection frequency may therefore be adjusted according to donor characteristics and processing requirements rather than according to a universal requirement for prolonged abstinence.
For scientific studies in which sperm count or total sperm yield is not the primary endpoint—such as investigations of motility, sperm DNA integrity, oxidative stress, mitochondrial function, proteomics, transcriptomics, or other functional outcomes—prolonged abstinence is not necessary. Instead, the abstinence interval should be prespecified, narrowly standardized, applied consistently across study groups, and reported exactly. In such studies, minimizing pre-analytical variability is more important than maximizing semen volume or total sperm count.
Implications for ART outcomes
Although this meta-analysis primarily quantifies laboratory semen parameters, a subset of included studies reported assisted reproductive technology (ART) outcomes, including fertilization rate, embryo quality, implantation, clinical pregnancy, and live birth. These outcomes were not suitable for quantitative synthesis because of heterogeneous reporting formats, differing denominators, variable stimulation protocols, and limited extractable data.
Narratively, several ART-based studies suggest that shorter ejaculatory abstinence or consecutive ejaculation protocols may be associated with improved fertilization or embryo development, particularly in men with poor baseline semen quality or elevated sperm DNA fragmentation. These findings are biologically plausible given the consistent improvement in sperm motility and DNA integrity observed with shorter abstinence in the present meta-analysis. However, other studies report neutral or inconsistent effects, and few provide adequately powered comparisons for live birth.
Consequently, while laboratory improvements in motility and DNA integrity may support the use of shorter abstinence in selected ART contexts, firm conclusions regarding fertilization, pregnancy, or live birth outcomes cannot be drawn from the current evidence base. Dedicated trials with standardized abstinence definitions and reproductive endpoints are required before laboratory-based associations can be confidently translated into clinical ART recommendations.
Strengths and limitations
The strengths of this study include its prospective protocol registration, adherence to PRISMA 2020 standards, comprehensive database searching, detailed subgroup analyses, and robust sensitivity analyses. The inclusion of diverse populations and study designs enhances generalizability.
However, several limitations should be acknowledged. First, substantial heterogeneity was observed across all semen parameters, reflecting differences in study design, abstinence categorization, laboratory methods, and participant characteristics. Second, most included studies were observational, and residual confounding cannot be excluded. Third, paired-ejaculate studies often lacked reported correlation coefficients, necessitating conservative analytical assumptions. Finally, publication bias or small-study effects were suggested for some outcomes, although their impact on motility and morphology appeared limited.
A key limitation of this meta-analysis is the very high between-study heterogeneity observed across nearly all pooled outcomes (I² frequently > 85–95%). This heterogeneity reflects substantial variability in study populations (fertile versus subfertile or ART cohorts), definitions and categorization of ejaculatory abstinence (ranging from ≤ 4 h to ≥ 14 days), semen analysis techniques (manual microscopy versus CASA), and the use of different World Health Organization laboratory manual editions over time. Despite the use of random-effects models, subgroup analyses, and extensive sensitivity analyses, a large proportion of this heterogeneity remained unexplained. Consequently, pooled effect estimates should be interpreted as average trends across heterogeneous contexts rather than definitive estimates for individual patients. This high inconsistency directly informed the GRADE downgrading for inconsistency and limits the certainty and generalizability of pooled estimates, supporting individualized rather than universal abstinence recommendations. In addition, temporal variation in semen analysis standards across successive WHO manual editions and differences between manual and CASA-based assessments represent important sources of measurement heterogeneity.
Another limitation is the coexistence of different experimental frameworks in the literature on ejaculatory abstinence. Some studies evaluate semen parameters after different abstinence intervals before a single ejaculation, whereas others investigate repeated or consecutive ejaculations. Although these approaches are biologically related, they represent distinct research questions and therefore required separate interpretation in the present review.
Implications for future research
Future research should prioritize well-designed randomized or crossover trials directly comparing abstinence intervals with standardized outcome reporting, including sperm DNA integrity and live birth outcomes. Studies incorporating time-to-pregnancy, fertility outcomes, and patient-centered endpoints are needed to translate laboratory findings into clinical recommendations. Additionally, mechanistic studies exploring oxidative stress, epididymal aging, and molecular sperm markers may further clarify abstinence-related effects.
Conclusions
This systematic review found that ejaculatory abstinence duration was associated with different patterns in laboratory semen parameters. Higher-duration intervals were associated with greater semen volume and total sperm count, whereas lower-duration intervals were associated with higher total and progressive sperm motility. No consistent association was identified for sperm concentration or morphology. These findings were based on low- or very-low-certainty evidence and should not be interpreted as causal or as defining a universally optimal abstinence interval. Repeated or frequent ejaculation protocols and reproductive outcomes were synthesized separately and were not considered equivalent to discrete abstinence-interval comparisons.
These findings support a personalized approach to abstinence recommendations, taking into account the clinical context, fertility goals, and underlying semen characteristics. Shorter abstinence intervals may be advantageous when sperm functional quality is prioritized, particularly in subfertile men and ART settings, while longer abstinence may be appropriate when maximizing sperm quantity is the primary objective. However, the very high heterogeneity across studies limits the certainty and direct transportability of pooled effect sizes, reinforcing that these results represent average laboratory associations rather than universally applicable clinical thresholds. Importantly, because reproductive outcomes were not quantitatively synthesized, these conclusions should not be interpreted as direct evidence of improved fertilization, pregnancy, or live birth, but rather as laboratory associations that may inform individualized clinical decision-making.
Future adequately powered randomized or crossover studies using standardized ejaculatory abstinence intervals and consistent laboratory methods are needed to confirm these associations. Such studies should prioritize reporting of sperm DNA integrity and reproductive endpoints, including fertilization, clinical pregnancy, and live birth, to enable clinically decisive recommendations.
Supplementary Information
Supplementary Material 1: Table S1. Characteristics of included studies. Table S2. Advanced sperm function outcomes according to ejaculatory abstinence duration.
Acknowledgements
AI use disclosure: Generative artificial intelligence tools, including ChatGPT by OpenAI, were used to assist with language editing, improvement of readability, organization of selected sections, and formatting of tables and figure captions. These tools were not used to generate or alter study data, perform statistical analyses, select studies, assess risk of bias, or determine scientific conclusions. All AI-assisted content was critically reviewed, verified, and revised by the authors, who take full responsibility for the accuracy, integrity, and final content of the manuscript.
Institutional review board statement
Not applicable.
Informed consent statement
Not applicable.
Authors’ contributions
Conceptualization: Bakhtiyar Serik, Riza Sharapatova, Amin TamadonMethodology: Bakhtiyar Serik, Riza Sharapatova, Amin Tamadon, Akmaral BaspakovaSoftware: Bakhtiyar Serik, Amin TamadonValidation: Bakhtiyar Serik, Riza Sharapatova, Akylbek Pshenbayev, Aizhan Ormankyzy, Kulyash R. ZhilisbayevaFormal analysis: Bakhtiyar Serik, Amin Tamadon, Nadiar M. MussinInvestigation: Riza Sharapatova, Amin Tamadon, Akmaral BaspakovaResources: Nadiar M. Mussin, Akylbek PshenbayevData curation: Bakhtiyar Serik, Akylbek Pshenbayev, Aizhan Ormankyzy, Kulyash R. ZhilisbayevaWriting – original draft: Riza Sharapatova, Amin TamadonWriting – review & editing: Bakhtiyar Serik, Nadiar M. Mussin, Kulyash R. Zhilisbayeva, Akmaral Baspakova, Akylbek Pshenbayev, Aizhan OrmankyzyVisualization: Bakhtiyar Serik, Amin TamadonSupervision: Bakhtiyar Serik, Amin TamadonProject administration: Bakhtiyar Serik, Amin TamadonFunding acquisition: Not applicableAll authors have read and approved the final version of the manuscript and agree to be accountable for all aspects of the work.
Funding
This research was funded by the authors.
Data availability
The data presented in this study are available from the corresponding authors upon reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Riza Sharapatova, Email: rsharapatova@gmail.com.
Amin Tamadon, Email: amintamaddon@yahoo.com.
References
- 1.Abdul Rahman H, Mazlan B, Amdani SN. A soping review of male fertility assessment and treatment across historical and contemporary paradigms. ANDROLOGY. 2025;n/a(n/a):e70156. 10.1111/andr.70156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Pelzman DL, Sandlow JI. Sperm morphology: Evaluating its clinical relevance in contemporary fertility practice. Reprod Med Biol. 2024;23(1):e12594. 10.1002/rmb2.12594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Russo GI, Asmundo MG, Cocci A, Abdelhameed AS, Liprino A, Giacone F, et al. Ejaculatory abstinence duration impacts semen parameters: Insights from a retrospective analysis in male infertility on 23,527 analyses. Front Endocrinol. 2025;16:1529262. 10.3389/fendo.2025.1529262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bjorndahl L, Kirkman Brown J, other Editorial Board Members of the WHOLMftE. Processing of Human S. The sixth edition of the WHO Laboratory Manual for the Examination and Processing of Human Semen: ensuring quality and standardization in basic examination of human ejaculates. Fertil Steril. 2022;117(2):246–51. 10.1016/j.fertnstert.2021.12.012. [DOI] [PubMed] [Google Scholar]
- 5.Xie M, Hammerli S, Leeners B. The association between abstinence period and semen parameters in humans: results in normal samples and different sperm pathology. Life. 2024;14(2):188. 10.3390/life14020188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lo Giudice A, Asmundo MG, Cimino S, Cocci A, Falcone M, Capece M, et al. Effects of long and short ejaculatory abstinence on sperm parameters: a meta-analysis of randomized-controlled trials. Front Endocrinol. 2024;15:1373426. 10.3389/fendo.2024.1373426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Du C, Li Y, Yin C, Luo X, Pan X. Association of abstinence time with semen quality and fertility outcomes: a systematic review and dose-response meta-analysis. ANDROLOGY. 2024;12(6):1224–35. 10.1111/andr.13583. [DOI] [PubMed] [Google Scholar]
- 8.Raditya M, Hari Soejono A, Siswanto MA, Atmoko W, Shah R, Agarwal A, et al. Impact of shorter abstinence periods on semen parameters: A systematic review and meta-analysis. World J Mens Health. 2025;43(3):563–79. 10.5534/wjmh.240035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tanga BM, Qamar AY, Raza S, Bang S, Fang X, Yoon K, et al. Semen evaluation: methodological advancements in sperm quality-specific fertility assessment - A review. Anim Biosci. 2021;34(8):1253–70. 10.5713/ab.21.0072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hanson BM, Aston KI, Jenkins TG, Carrell DT, Hotaling JM. The impact of ejaculatory abstinence on semen analysis parameters: a systematic review. J Assist Reprod Genet. 2018;35(2):213–20. 10.1007/s10815-017-1086-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Mayorga-Torres JM, Agarwal A, Roychoudhury S, Cadavid A, Cardona-Maya WD. Can a Short Term of Repeated Ejaculations Affect Seminal Parameters? J Reprod Infertil. 2016;17(3):177–83. [PMC free article] [PubMed] [Google Scholar]
- 12.Halim B, Angellee J, Agustinus A, Pakpahan C, Lubis HP. Enhancing sperm quality through consecutive ejaculation after short abstinence in men with low semen parameters undergoing ICSI. J Reprod Infertil. 2024;25(4):281–9. 10.18502/jri.v25i4.18126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wang Y, Fu X, Li H. Mechanisms of oxidative stress-induced sperm dysfunction. Front Endocrinol. 2025;16:1520835. 10.3389/fendo.2025.1520835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sengupta P, Dutta S. Repainting the old fence: connecting curiosity and conception on semen retention, ejaculatory abstinence, and sperm metabolism. Middle East Fertil Soc J. 2025;30(1):16. 10.1186/s43043-025-00229-6. [DOI] [Google Scholar]
- 15.Poopaibool N, Tangprasittipap A, Chumchuen S, Satirapod C, Singwongsa A. Effects of a short abstinence period on sperm quality in oligozoospermic men. Clin Exp Reprod Med. 2023;50(4):262–9. 10.5653/cerm.2023.06100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chen GX, Li HY, Lin YH, Huang ZQ, Huang PY, Da LC, et al. The effect of age and abstinence time on semen quality: a retrospective study. Asian J Androl. 2022;24(1):73–7. 10.4103/aja202165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chung E, Atmoko W, Saleh R, Shah R, Agarwal A. Sixth edition of the World Health Organization laboratory manual of semen analysis: Updates and essential take away for busy clinicians. Arab J Urol. 2024;22(2):71–4. 10.1080/20905998.2023.2298048. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1: Table S1. Characteristics of included studies. Table S2. Advanced sperm function outcomes according to ejaculatory abstinence duration.
Data Availability Statement
The data presented in this study are available from the corresponding authors upon reasonable request.