Seminal plasma cytokines in fertile versus infertile men: a systematic review and meta-analysis.

Clinical analysis of male fertility is an area of increasing interest, with mounting evidence for declining sperm counts and other measures of male reproductive health over recent decades ( Levine et al. , 2017 ; Garcia-Grau et al. , 2022 ; Luo et al. , 2023 ). To date, assessment of male fertility has relied upon conventional semen analysis, wherein sperm count, motility, morphology, semen volume, and viscosity are evaluated ( World Health Organization, 2021 ). However, having normal sperm parameters does not assure fertility in men ( Sakkas et al. , 2015 ), and the pathophysiological origins of impaired male fertility are not often clear ( Aitken et al. , 2010 ; Gatimel et al. , 2017 ). While semen analysis will remain the cornerstone of fertility testing, improved prognostic value and greater insight into the underlying causes might be afforded by additional measures of male reproductive tract function ( World Health Organization, 2021 ). This may be particularly relevant in cases of unexplained and idiopathic infertility, where standard semen analysis falls short. One possibility is the analysis of cytokines and chemokines in seminal plasma (SP). Indeed, the sixth edition of the WHO laboratory manual for the examination and processing of human semen suggests that evaluating SP cytokines may help identify men with genital tract infection and/or inflammatory conditions of the reproductive tract. Cytokines comprise a large and divergent range of low molecular weight proteins with multifaceted roles in maintaining tissue homeostasis that span immune regulation, control of infection and inflammation, and tissue development, remodeling, and repair. In the context of male reproduction, cytokines play a number of important functions, including supporting spermatogenesis, mediating immune tolerance and quality control of gametes, and homeostasis and secretory function of the male accessory organs ( Loveland et al. , 2017 ). SP cytokines have an additional role in promoting female reproductive investment after seminal fluid intromission ( Schjenken and Robertson, 2020 ). Depending on their identity and the balance of factors, cytokines can exert both positive and negative effects on sperm survival, fertilisation potential, and developmental competence ( Seshadri et al. , 2009 ; Fraczek and Kurpisz, 2015 ; Loveland et al. , 2017 ), and also interact with the female reproductive tract to induce an immune response conducive to conception and pregnancy ( Robertson and Sharkey, 2016 ; Schjenken and Robertson, 2020 ). There is an emerging view that the profile of cytokines and other soluble immune-regulatory mediators in SP may reflect the immunological or inflammatory state of the tissues from which those fluids originate ( Fraczek and Kurpisz, 2015 ; Loveland et al. , 2017 ; Dutta et al. , 2025 ; Ma et al. , 2025 ). As in other secretory tissues and mucosal surfaces, their relative abundance is expected to be responsive to a range of local and systemic physiological and environmental cues and clinical conditions. Studies investigating priorities in andrology have ranked male infertility, prostatitis, prostate inflammation, varicocele, and male accessory gland infection as amongst the top 10 research themes ( Crafa et al. , 2023 ). Many of these conditions involve acute or chronic inflammation and immune activation in male reproductive tissues, and are associated with elevated abundance of pro-inflammatory cytokines in SP ( Orhan et al. , 2001 ; Penna et al. , 2007 ; Fraczek and Kurpisz, 2015 ; Zeinali et al. , 2017 ; Lyons et al. , 2023 ). There is a strong biological rationale and emerging evidence that fertility status in men can be affected by both acute and chronic inflammation in the testes and male accessory organs, although the pathophysiological impact on semen parameters and male fertility is not yet certain ( Dutta et al. , 2025 ; Ma et al. , 2025 ). Therefore, gaining a comprehensive understanding of the significance of the cytokine profile in SP and its association with inflammatory conditions and male fertility is an important objective in male reproductive health research. To date, a substantial research effort has focused on determining the identity and abundance of cytokines and chemokines within male SP and investigating whether their concentration differs between fertile and infertile men. Seminal plasma composition, and the relative abundance of cytokines it contains are also being investigated for their potential in predicting successful pregnancy outcomes in assisted conception settings. Many studies have sought to explore potential relationships between cytokine abundance in SP and measures of semen quality, including sperm concentration ( Paradisi et al. , 1997 ; Furuya et al. , 2003 ; Sanocka et al. , 2003 ; Matalliotakis et al. , 2006 ), motility ( Gruschwitz et al. , 1996 ; Paradisi et al. , 1997 ; Kocak et al. , 2002 ; Sanocka et al. , 2003 ; Matalliotakis et al. , 2006 ), morphology ( Paradisi et al. , 1997 ; Furuya et al. , 2003 ), and viscosity ( Castiglione et al. , 2014 ). However, often contradictory findings are reported between studies. We have recently reported on the limitations inherent in the published literature and considerations for improving studies on seminal plasma cytokine content ( Lyons et al. , 2023 ). Critically, a lack of standardized guidelines for cytokine quantification in SP, coupled with a lack of established reference ranges (or decision limits) for various cytokines and chemokines in healthy fertile men, limits the capacity to compare findings across studies. Seminal plasma cytokine concentrations have also been shown to fluctuate both between and within individual men over time ( Sharkey et al. , 2016 , 2017 ). This makes it difficult for the clinical utility of measuring SP cytokines to be realised, as without reference ranges from healthy proven fertile men to compare with, studies are limited by technical differences and variation in factors such as the criteria used to define control and infertile cohorts. Notwithstanding these limitations, a substantial body of work appears to indicate differences in SP cytokine abundance between men with differing clinical and fertility status. The purpose of this systematic review was to assemble a synthesis of existing studies to determine whether collectively these studies support an association between cytokine abundance in SP and fertility status. The overarching aim is to identify cytokines that might be associated with fertility status in men, and that have potential as biomarkers that ultimately might be utilised to improve clinical assessment of male fertility in couples seeking assisted conception.

This systematic review was registered with the International Prospective Register of Systematic Reviews (PROSPERO) and assigned the registration number CRD42023398438. Our systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) ( Moher et al. , 2015 ) and the Meta-analysis for Observational Studies in Epidemiology (MOOSE) reporting statement ( Stroup et al. , 2000 ). A systematic search of the scientific literature via databases including PubMed, Web of Science, Scopus, and Embase was conducted to identify studies that had quantified cytokines and chemokines in SP. The search terms combined keywords relating to semen, including seminal plasma, seminal fluid, semen, and ejaculate, and cytokine-associated keywords, including interleukin, growth factor, cytokine, chemokine, monokine, lymphokine, interferon, tumour necrosis factor, and human leukocyte antigen. Studies were restricted to those including human participants by using search terms ‘human’, ‘men’, ‘donor’, ‘participant’, and ‘patient’. The full search strategy is provided in Supplementary Table S1 . Databases were searched from publication dates spanning inception until April 2025 inclusive to identify all relevant papers. For this study, ‘healthy control’ cohorts were defined as comprising either proven fertile men (with a previous spontaneous conception with a female partner) or normozoospermic men (generally healthy with normal sperm parameters by semen analysis and no history of infertility). ‘Infertile’ cohorts were defined as comprising men with either male factor infertility (oligo-, astheno-, terato-, and azoo- spermia, or any combination of these), or idiopathic infertility (also termed ‘unexplained infertility’, with no spontaneous conception following ≥12 months of regular unprotected intercourse, despite normal sperm parameters). The literature search was limited to articles written in the English language and comprising human subjects. Studies that did not contain full original datasets, including reviews, conference abstracts, theses/dissertations, opinion pieces, and editorial letters, were excluded. Any studies that did not quantify cytokine or chemokine concentration in SP samples, reported results inappropriately (i.e. without clearly legible axes) or were intervention studies (i.e. included an exercise/diet regime, etc.) were excluded. To meet eligibility, studies were required to report clear definitions of both the control and experimental groups. Studies that did not clearly report the fertility status of cohorts, or that did not have clearly discernible datasets for at least one cohort meeting the above definition of ‘healthy control’ and at least one cohort meeting the definition of ‘infertile’, were excluded. Studies were excluded if relevant cohorts included men with chronic medical conditions and diseases, men with significant smoking habits, a history of recent drug and/or alcohol abuse, or men with vasectomy. Studies where one or more infertile cohort was further defined by an inflammatory disease, including leukocytospermia, urogenital disease, infection, varicocele, or hydrocele were included, but only if the cohorts also met the above definition of infertility. In this event, the associated diseases were recorded and factored into the analysis. Finally, studies that did not evaluate SP but instead evaluated only whole semen or spermatozoa-associated (membrane-bound) factors were excluded. All articles were imported directly from the scientific literature databases to Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org ). Following the removal of duplicates (automatic and manual removals), one reviewer (H.E.L) screened titles and abstracts, then completed a full text review of the studies. Queries relating to the appropriateness of including specific studies were resolved in consultation with additional reviewers (S.A.R and D.J.S). Two reviewers (H.E.L and C.N.AS) performed data extraction on included studies, extracting study demographics (title, authors, country of origin), participant demographics and baseline characteristics (number of participants, age, period of abstinence, semen analysis results, and fertility status), and outcome data (study design, method for processing of SP, platform used for cytokine quantification, and reported cytokine concentrations). Data extraction was limited to individual study summary statistics and did not include data files on individual patients in each study, due to limited data availability. H.E.L and C.N.AS analysed the extracted data and synthesised the results. A PRISMA flow diagram outlining the study selection, number of studies included within the review and reasons for exclusion are included in Fig. 1 . Cytokine and chemokine concentration data extracted from the 48 studies are provided in Supplementary Table S2 . Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram . Adapted from Moher et al . (2015) . A quality assessment was performed on all studies that underwent data extraction by two reviewers (H.E.L and C.N.AS) using the NHLBI-NIH quality assessment tool for observation cohort and cross-sectional studies ( https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools ). Studies were scored according to alignment with the structured questions criteria, with a single point assigned per question and a maximum score of 10 points. The scoring corresponded to assessment as good (scores of 7–10), fair (3–6), and poor (0–3). Studies scoring less than 3 points were deemed to be of poor quality, and those containing multiple “fatal flaws” were excluded from the final dataset. Factors taken into consideration included risk of bias due to reporting, missing cohort descriptions such as n-values, ages, period of recruitment, and in particular, factors known to impact cytokine concentration, such as period of abstinence, BMI, substance abuse, and smoking status. Any discrepancies between the reviewers were resolved by S.A.R and D.J.S. All effect measures were unadjusted for confounders. A qualitative synthesis was performed for all parameters included in this review. Seminal plasma factors were categorised as either cytokines, chemokines, or growth factors. These were then further subcategorised into the individual families to which the specific factor belonged, including interleukins (ILs), colony-stimulating factors (CSFs), transforming growth factor beta (TGFB) superfamily members, C-X-C motif chemokine ligands, CC chemokine ligands, interferons (IFNs), and tumour necrosis factors (TNFs). A meta-analysis was performed for parameters where five or more studies provided data on cytokine concentrations in a control fertile/normozoospermic population and a defined infertile population. Only four cytokines (IL1B, IL6, TNFA, and CXCL8) satisfied these criteria. To be included in the meta-analysis, studies had to report data in similar/identical descriptive statistics, as whole datasets were not available. Studies that presented results as mean ± SD were directly included in the meta-analysis. When datasets were presented as mean ± SEM, SEM values were first transformed to SD according to the methods described in the Cochrane handbook ( Higgins et al. , 2019 ). Datasets presented as either median and range or median and IQR were excluded from meta-analyses. In circumstances where infertile cohorts were sub-classified based on sperm defects (i.e. oligozoospermia, asthenozoospermia etc.) instead of being presented as a single infertile cohort, individual datasets were combined to form a single infertile dataset according to recommendations provided by the Cochrane handbook ( Higgins et al. , 2019 ) and Introduction to Meta-Analysis ( Borenstein et al. , 2009 ) resource. Meta-analyses were performed using R (version 4.2.2) and RStudio (version 2022.12.0), and the metafor package (a meta-analysis package for R; version 4.2-0) ( Viechtbauer, 2010 ), in accordance with the method described by Quintana (2015) . Effect sizes were calculated for a two-group comparison design using the standardized mean differences (SMD), and then an inverse variance-weighted random-effects model was fitted to the effect sizes. This allowed for the interpretation of low ( d  = 0.2), medium ( d  = 0.5), and large ( d  = 0.8) effect sizes. To estimate heterogeneity, the restricted maximum-likelihood (REML) estimator was used ( Patterson and Thompson, 1971 ). In addition, heterogeneity analysis was performed using the Q -test for heterogeneity ( Cochran, 1954 ), and the I 2 statistics ( Higgins and Thompson, 2002 ; Higgins et al. , 2003 ) were reported. Tau 2 and tau were also calculated but were not reported on the forest plots. To detect whether any moderator effects existed due to factors including control group description (i.e. fertile or normozoospermic) or paper quality (good, fair, low) an omnibus test of model coefficients was calculated, to provide a QM and P -value for tested moderators ( Viechtbauer, 2010 ). To detect whether any studies could be considered overly influential on meta-analysis results, studentized residuals ( Pope, 1976 ) and Cook’s distances ( Cook, 1982 ) were evaluated for statistical significance. When a study was identified as being overly influential, a sensitivity analysis was performed. To detect the impact of small studies on effect size and identify potential publication bias, funnel plots were generated when enough studies were present to allow this, and then Egger’s regression ( Egger et al. , 1997 ), the weight-function model ( Vevea and Hedges, 1995 ), and the rank correlation test ( Begg and Mazumdar, 1994 ) were calculated to assess asymmetry within the funnel plot. GraphPad Prism (version 10.1.1, Boston, Massachusetts, USA) was used to generate bubble plots.

A total of 5737 references were retrieved (PubMed: 1273, Scopus: 881, Web of Science: 1538, and Embase: 2045). Of these, 2737 studies were removed as duplicate articles, with 3000 articles remaining for title and abstract screening. A further 2709 articles were deemed ineligible and removed during this phase, leaving 291 articles for full-text screening. A further 220 articles were subsequently excluded, most commonly because of inappropriate classification of patient populations. Seventy-one articles then underwent full data extraction and quality assessment, with a further 19 articles removed following quality assessment, leaving a total of 52 studies eligible for inclusion. The proportion of studies identified as good, fair, and poor is outlined in Supplementary Fig. S1 . Study cohort descriptions are summarised in Table 1 . The proportion of included studies reporting healthy control populations as either fertile or normozoospermic were 56% and 44%, respectively ( Supplementary Fig. S1 ). The proportion of infertile cohorts categorised as male factor, idiopathic, or inflammatory disease-associated infertility were 49%, 19% and 32%, respectively ( Supplementary Fig. S1 ). The types of infertility examined in different studies was collected during data extraction and are indicated in Table 1 . Participant characteristics for included studies. 45 total 15 control 30 infertile [15 O (5 also have V) 15 A] Oligozoospermic Azoospermic Varicocele C: 34.0 ± 5.2 O: 33.0 ± 4.8 A: 37.0 ± 5.4 175 total (2 HPV, 15 B) 31 control 144 infertile Primary male factor Human papilloma virus C. trachomatis C: 32.0 (23.0–46.0) IF: 32.0 (18.0–50.0) 51 total 14 control 37 infertile IL6 TNFA 70 total 25 control 45 infertile IL6 IL10 PGE2 211 total 42 control 74 BP 95 BBP Oligoasthenoteratozoospermia male accessory gland infection - bacterial prostatitis and bilateral bacterial prostatitis C: 35.1 (24.0–40.0) IF: 34.6 (27.0–41.0) TNFA IL6 IL10 346 total 103 control 243 infertile (152 O 142 AST 90 OAST) Oligozoospermia Asthenozoospermia Oligoasthenozoospermia C: 33.0 ± 6.0 O: 34.0 ± 5.0 A: 34.0 ± 6.0 OAST: 34.0 ± 5.0 IL1B IL6 IL12 CXCL8 TNFA MCP1 MIP1B 140 total 21 control 119 infertile (33 INF) IL1B IL2 IL6 sIL2R sIL6R 92 total 12 control 32 INFLAM 48 NON-INFLAM IL4 IL6 IL17 IL23 IFNG TRAIL 91 total 15 control 44 AST-T 20 OAT 12 A Primary male factor Asthenoteratozoospermia Oligoasthenoteratozoospermia Azoospermia IL1B IL18 91 total 15 control 44 AST-T 20 OAT 12 A Primary male factor Asthenoteratozoospermia Oligoasthenoteratozoospermia Azoospermia IL6 CXCL8 120 total 38 control 38 FV 44 IF-V C: 35.0 ± 4.6 F-V: 34.0 ± 6.9 V: 33.3 ± 4.1 100 total 25 control 25 NORMO-L 25 IF-L 25 IF C: 28.7 ± 10.5 N-L: 28.2 ± 7.0 IF-L: 29.0 ± 6.9 IF: 27.8 ± 7.2 IL6 TNFA 70 total 25 control 15 T 15 AT 15 OAT Infertile men with primary infertility (Teratozoospermia, Asthenoteratozoospermia, or Oligoasthenoteratozoospermia) Normozoospermic generally healthy C: 33.64 ± 1.26 IF: 34.74 ± 0.70 IL10 IL18 200 total 100 control 100 V (50 OAT, 50 N) C: 31.6 ± 4.9 (22.0–47.0) OAT: 31.5 ± 5.2 (23.0–46.0) N: 31.7 ± 4.5 (22.0–40.0) 390 total 50 control 340 IF (218 INF) C: 36.5 (24.0–47.0) IF: 37.5 (18.0–59.0) 63 total 23 control 15 TSP 6 EO 19 COMBINED 290 total 109 control 181 (IF) A consecutive series of 699 male patients attending our outpatient clinic for the first time from September 2012 to September 2017 and seeking medical care for couple infertility. Fertile men were defined as partners of women in the second or third trimester of pregnancy or who had fathered a child during the previous year by natural conception. Out of 699 males of infertile couples, the MAR test was not assessable in 518 hence 181 patients were included. The MAR test was assessable in the entire fertile cohort. Control: 36.6 ± 5.2 IF: 37.0 ± 7.9 60 total 20 control 20 O-AS 20 A Oligoasthenozoospermic Azoospermic IL2 sIL2R 93 total 21 control 40 N 53 abnormal (17 AST 25 O-AST 11 A) Asthenozoospermia Oligoasthenozoospermia Azoospermia Cryptorchidism Mumps orchitis Idiopathic testicular lesion Klinefelter’s syndrome 82 total 19 control 17 V 17 INF 7 C 7 M 10 ITL 5 K Varicocele Infection Cryptorchidism Mumps orchitis Idiopathic testicular lesion Klinefelter’s syndrome 90 total 20 control 20 V-OAT 20 INF-OAT 20 ITL-OAT 10 A Oligoasthenoteratozoospermia Varicocele Infection Idiopathic testicular lesion Azoospermia IL2 IL6 IL11 CXCL8 80 total 18 control 62 infertile (17 INF 15 V 7 C 7 M 10 ITL 6 K) Varicocele Infection Cryptorchidism Mumps orchitis Idiopathic testicular lesion Klinefelter’s syndrome IL18 IFNG 70 total 14 control 32 L 24 V Leukocytospermia Varicocele C: 25.0–35.0 L: 28.0–38.0 V: 27.0–38.0 IL6 TNFA 53 total 12 control 17 V 14 INF 10 infertile Varicocele Genitourinary infections Inability to conceive >12 months C: 27–34 IF: 20–45 57 total 11 control 16 V 17 INF 13 Idio Varicocele Genitourinary infections Idiopathic infertility Inability to conceive >24 months C: 28–36 IF: 28–39 81 total 30 control 30 Immuno 21 IF 74 total 33 control 27 IF 14 Immuno 48 total 15 control 16 INF 17 no-INF 84 total 28 control 21 Immuno 35 IF (3 O 5 AST 11 T 16 MULTI) Male-factor infertile Immunoinfertile Oligozoospermia Asthenozoospermia Teratozoospermia Multiple sperm factors 131 total 25 control 106 IF HLA-G TGFB1 TGFB2 TGFB3 200 total 24 control 176 infertile C: 34.0 ± 8.0 (21.0–65.0) IF: 33.0 ± 9.0 (21.0–65.0) TNFA IL4 43 total 20 control 23 IF (21 OAT, 2 A, 2 L) 41 total 20 control 21 IF (19 OAT, 2 A, 2 L) C: 23.0–43.0 IF: 27.0–43.0 38 total 18 control 20 IF (15 V) C: 29.0–48.0 IF: 23.0–48.0 sIL2R IL4 IL6 114 total 20 control 32 (T) 33 (AT) 29 (OAT) Teratozoospermia Asthenoteratozoospermia Oligoasthenoteratozoospermia C: 29 (25–41) T: 32 (29–34) AT: 34 (30–36) OAT: 33 (30–35) IL1A IL4 IL6 IL18 CXCL8 GMCSF IFNG 57 total 33 control 24 IF IL6 IL23 CXCL8 TNFA 32 total 10 control 12 IF 10 INF IL2 CXCL8 52 total 11 control 41 INFLAM IL10 CXCL1 52 total 12 control 10 A 10 O 10 AST 10 V Azoospermia Oligozoospermia Asthenozoospermia Varicocele C: 33.0 ± 2.0 A: 33.0 ± 3.0 O: 33.0 ± 3.0 AST: 32.0 ± 1.0 V: 36.0 ± 4.0 73 total 15 control 28 A 30 O (15 V) Azoospermia Oligozoospermia Varicocele IL6 CXCL8 TNFA 61 total 22 control 39 INF IL1B IL6 CXCL8 TNFA 155 total 95 control 60 GTI IL1B IL6 CXCL8 TNFA 74 total 14 control 8 AST 13 O 19 O-AST 20 A Asthenozoospermia Oligozoospermia Oligoasthenozoospermia Azoospermia IL6 IL10 IL11 IL12 CXCL8 TNFA IFNG 86 total 20 control 30 I 36 AST Idiopathic infertility Asthenozoospermia IL1B IL6 sIL2R 69 total 9 control 44 IF 16 L IL6 CXCL8 MCP1 68 total 14 control 44 IF 10 L 1040 total 260 control 260 O 260 I 260 AST Oligozoospermia Idiopathic infertility Asthenozoospermia IL6 TNFA 42 total 19 control 23 IF Idiopathic infertility Asthenozoospermia Teratozoospermia Asthenoteratozoospermia Oligoasthenoteratozoospermia 20 total 6 control 7 V 7 INF Varicocele Genitourinary infection C: 31–40 IF: 29–40 2248 total 248 control 2000 IF C: 31.8 (20.0–50.0) IF: 32.5 (18.0–49.0) 62 total 34 control 28 IF C: 36.8 IF: 37.3 IL1B IL6 CXCL8 VEGF TNFA TGFB1 CSF1 71 total 16 control 55 IF C: 33.2 ± 6.0 IF: 33.5 ± 6.8 IL4 CCL11 A, azoospermia; AT, asthenoteratospermia; AST, asthenozoospermia; AST-T, asthenoteratozoospermia; B, bacteria present; BBP, bacterial bilateral prostatitis; BP, bacterial prostatitis; C, cryptorchidism; COMBINED, testicular secretory pathology and epididymal occlusion; EO, epididymal occlusion; F, fertile; FSH, follicle-stimulating hormone; GTI, genital tract infection; HPF, high-powered fields; HPV, human papilloma virus; I, idiopathic; IF, infertile; IMMUNO, immunoinfertile; INFLAM, inflammation; INF, infection; ITL, idiopathic testicular lesion; K, Klinefelter’s syndrome; L, leukocytes; LH, luteinising hormone; MAGI, male accessory gland infection; M, mumps; MF, male-factor; MULTI, multiple causes; NA, not assessed; ND, not determined; NORMO, normozoospermia; NR, not reported; ns, not significant; O, oligozoospermia; OAT, oligoasthenoteratozoospermia; OAST, oligoasthenozoospermia; T, teratozoospermia; TSP, testicular secretory pathology; V, varicocele; WBC, white blood cell. The systematic literature search identified 52 research articles, published between 1993 and April 2025, as meeting the eligibility criteria. All studies quantified the abundance of at least one cytokine or chemokine in SP, but many investigated several cytokines. Across the 52 studies, data on the reported concentration of 30 individual cytokines were extracted ( Supplemental Table S2 ). The most frequently investigated cytokines and chemokines were IL1B (n = 10 studies), IL6 (n = 23), TNFA (n = 13), and CXCL8 (n = 14). Data relating to the methods used for the processing and storage of SP in individual studies prior to cytokine assessment were also extracted ( Table 2 and Supplementary Table S2 ). In total, 16 different centrifugation conditions were identified, with 16 unique combinations of 15 speeds (200–16 000 × g), and 5 time settings (5–30 minutes). Of the 52 included studies, 40 studies reported storing aliquots of SP at temperatures between −18°C and −90°C, and 39 studies reported avoiding repeated freeze-thaw cycles. Most studies (39 of 52) reported using WHO guidelines to classify the fertility status of the male participant cohorts, utilising various editions from the first (1980) to the current sixth edition (2021). Reported seminal plasma processing methodology across studies. g, gravity; NR, not reported; RPM, revolutions per minute. A total of 20 studies were included for meta-analysis, with the aim of evaluating differences in SP cytokines in the infertile cohorts and the healthy control (fertile/normozoospermic) cohorts. The analysis included five studies on IL1B ( Supplementary Fig. S2A ) ( Shimonovitz et al. , 1994 ; Sanocka et al. , 2004 ; von Wolff et al. , 2007 ; Dziadecki et al. , 2010a ; Stabile et al. , 2023 ), 15 studies on IL6 ( Fig. 2A ) ( Shimonovitz et al. , 1994 ; Shimoya et al. , 1995 ; Paradisi et al. , 1997 ; Camejo et al. , 2001 ; Matalliotakis et al. , 2002 ; Camejo, 2003 ; Sanocka et al. , 2004 ; Kopa et al. , 2005 ; von Wolff et al. , 2007 ; Sakamoto et al. , 2008 ; Dziadecki et al. , 2010 ; Qian et al. , 2011 ; Shukla et al. , 2013 ; Eldamnhoury et al. , 2018 ; Micheli et al. , 2019 ), 9 studies on TNFA ( Fig. 2B ) ( Omu et al. , 1999 ; Camejo et al. , 2001 ; Sanocka et al. , 2004 ; von Wolff et al. , 2007 ; Sakamoto et al. , 2008 ; Qian et al. , 2011 ; Shukla et al. , 2013 ; Eldamnhoury et al. , 2018 ; Micheli et al. , 2019 ), and 9 studies on CXCL8 ( Fig. 2C ) ( Shimoya et al. , 1993 ; Rajasekaran et al. , 1995 ; Shimoya et al. , 1995 ; Matalliotakis et al. , 2002 ; Sanocka et al. , 2004 ; von Wolff et al. , 2007 ; Sakamoto et al. , 2008 ; Dziadecki et al. , 2010 ; Qian et al. , 2011 ). Meta-analysis forest plots of seminal plasma cytokine concentration between healthy control (fertile/normozoospermic) men and infertile men, regardless of type of infertility, representing standardized mean difference in IL6 (A), TNFA (B), and CXCL8 (IL8) (C) concentration . Data from individual studies are shown as box and whisker plots representing standardized mean effect size and 95% CI. Only data from studies reporting data as mean ± SD or mean ± SEM are included. Overall meta-analysis standardized mean effect size is shown as a green diamond representing the mean and 95% CI. Results are also shown numerically with P -value to indicate statistical significance. Quality represents the risk of bias as determined during quality assessment (poor, fair, good), the fertility status of the controls was stated as either ‘fertile’ (proven fertile) or ‘normo’ (normozoospermic), and weights describing the individual calculated weighting applied to each study for the calculation of the overall effect. Measures of heterogeneity are reported as Q score and P -value, as well as I 2 percentage. Effect sizes can be interpreted as weak (≤0.2), moderate (0.2–0.5), or strong (≥0.5). Firstly, we sought to investigate whether any differences in SP cytokine concentrations exist between the infertile and control cohorts, irrespective of the infertility condition, by performing a random-effects model meta-analysis. A significant effect of fertility status was observed for IL6 (n = 15, SMD [95% CI] = 0.39 [0.14–0.64], P  = 0.002; Fig. 2A ), TNFA (n = 9, SMD [95% CI] = 0.13 [0.00–0.25], P  = 0.042; Fig. 2B ), and CXCL8 (n = 9, SMD [95% CI] = 0.24 [0.06–0.43], P  = 0.01; Fig. 2C ), in favour of higher concentrations in the infertile cohorts. In contrast, meta-analysis of IL1B (n = 5, SMD [95% CI] = 0.00 [−0.54–0.55], P  = 0.99; Supplementary Fig. S2A ) showed no significant effect of fertility status. Measures of heterogeneity were also performed for all meta-analyses. IL1B and IL6 analyses showed I 2 scores indicative of high heterogeneity between studies ( I 2 = 77.8% and I 2 = 80.7%, respectively, both P <  0.001). Low levels of heterogeneity were observed for TNFA and CXCL8 ( I 2 = 4.3%, P  = 0.25, and I 2 = 0.0%, P  = 0.71 respectively). The influence of individual studies on the outcome of the meta-analysis was evaluated by calculating Cook’s distance and studentized residuals. No studies were identified as being an outlier and/or overly influential for IL6, or CXCL8, however one study was identified for IL1B ( Shimonovitz et al. , 1994 ). When this paper was removed from the IL1B meta-analysis, there was an increase in effect size favouring higher abundance in the healthy control men, but the difference did not meet statistical significance (n = 4, SMD [95% CI] = −0.19 [−0.71–0.32], P  = 0.07). Removal of this paper also moderately reduced the degree of heterogeneity amongst the IL1B studies (Q (df = 3) = 9.3, P  = 0.026, I 2 = 67.3%). However, with just four studies remaining, a valid conclusion cannot be drawn ( Supplementary Fig. S2B ). In the TNFA analysis, one study showed evidence of being an outlier ( Shukla et al. , 2013 ). Omission of this study from the TNFA meta-analysis did not substantially change the result, with no change to statistical significance and a slight increase in effect size (n = 8, SMD [95% CI] = 0.2 [0.02–0.39], P  = 0.034) ( Supplementary Fig. S3B ). To examine the significance of inflammatory conditions for SP cytokines, we next examined differences in SP cytokines between healthy control and infertile cohorts when the analysis was limited to studies wherein infertility was associated with a diagnosed inflammatory disease. A total of eight studies investigated IL6 in SP of cohorts fitting this criteria ( Shimoya et al. , 1995 ; Paradisi et al. , 1997 ; Matalliotakis et al. , 2002 ; Sanocka et al. , 2004 ; Kopa et al. , 2005 ; Sakamoto et al. , 2008 ; Eldamnhoury et al. , 2018 ; Micheli et al. , 2019 ), five studies investigated TNFA ( Omu et al. , 1999 ; Sanocka et al. , 2004 ; Sakamoto et al. , 2008 ; Eldamnhoury et al. , 2018 ; Micheli et al. , 2019 ), and six investigated CXCL8 ( Shimoya et al. , 1993 , 1995 ; Rajasekaran et al. , 1995 ; Matalliotakis et al. , 2002 ; Sanocka et al. , 2004 ; Sakamoto et al. , 2008 ). Meta-analysis of IL6 in studies where inflammatory disease was evident revealed a significant effect of fertility status on cytokine abundance in favour of higher concentrations in the infertile group (n = 8, SMD [95% CI] = 0.92 [0.29–1.55], P  = 0.004; Fig. 3A ). Cook’s distance and studentized residual analysis identified one study as being overly influential ( Shimoya et al. , 1995 ). When this study was removed from the analysis, the significant effect in favour of higher IL6 concentration in SP of infertile cohorts with inflammatory conditions remained (n = 7, SMD [95% CI] = 0.61 [0.27–0.95], P  = 0.011) ( Supplementary Fig. S4A ). Meta-analysis forest plots of seminal plasma cytokine concentration between healthy control (fertile/normozoospermic) men and men with infertility associated with inflammatory disease, representing standardized mean difference in IL6 (A), TNFA (B), and CXCL8 (IL8) (C) concentration . Data from individual studies are shown as box and whisker plots representing standardized mean effect size and 95% CI. Only data from studies reporting data as mean ± SD or mean ± SEM are included. Overall meta-analysis standardized mean effect size is shown as a green diamond representing the mean and 95% CI. Results are also shown numerically with P -value to indicate statistical significance. Quality represents the risk of bias as determined during quality assessment (poor, fair, good), fertility status of the controls was stated as either ‘fertile’ (proven fertile) or ‘normo’ (normozoospermic), and weights describing the individual calculated weighting applied to each study for the calculation of the overall effect. Measures of heterogeneity are reported as Q score and P -value, as well as I 2 percentage. Effect sizes can be interpreted as weak (≤0.2), moderate (0.2–0.5), or strong (≥0.5). Meta-analysis of TNFA in studies where inflammatory disease was evident showed a significant effect of fertility status in favour of higher concentrations in infertile cohorts (n = 5, SMD [95% CI] = 0.59 [0.19–0.98], P  = 0.004; Fig. 3B ). When one study identified as being overly influential ( Sanocka et al. , 2004 ) was removed from analysis, the effect size was strengthened and the level of heterogeneity reduced (n = 4, SMD [95% CI] = 0.78 [0.50–1.06], P  < 0.001) ( Supplementary Fig. S4B ). Meta-analysis of CXCL8 in studies where inflammatory disease was evident did not show an effect of fertility status (n = 6, SMD [95% CI] = 1.27 [−0.04 to 2.58], P  = 0.057; Fig. 3C ). Influence testing identified one study ( Rajasekaran et al. , 1995 ) as being overly influential, however its removal did not change the relationship with fertility status (n = 5, SMD [95% CI] = 0.68 [−0.11 to 1.46], P  = 0.091) ( Supplementary Fig. S4C ). When assessing the influence of bias on the IL6, TNFA, and CXCL8 meta-analysis for small study and publication bias, we observed no significant funnel plot asymmetry (IL6: z  = 1.24, P  = 0.22 and Kendall’s tau = 0.30, P  = 0.14; TNFA: z  = 1.44, P  = 0.15 and Kendall’s tau = 0.50, P  = 0.08; and CXCL8: z  = 1.35, P  = 0.18 and Kendall’s tau = 0.39, P  = 0.18), suggesting there is no effect of publication and small sample size bias across studies ( Fig. 4 ). Assessment of small study, and publication bias within the meta-analysis evaluating differences in seminal plasma IL6 (A), TNFA (B), and CXCL8 (IL8) (C) concentrations between fertile/normozoospermic cohorts and combined infertile cohorts . Egger’s regression test ( z -score and P -value) and rank correlation test (Kendall’s tau and P -value) as objective tests of funnel plot asymmetry. A P- value <0.05 is considered statistically significant. Moderator effects also potentially contribute to the observed heterogeneity. To investigate this, we evaluated the impact of paper quality and control fertility status, but no effect across these cytokines and chemokines was found. Across the 52 studies, 11 members of the interleukin superfamily were investigated, as well as two receptor antagonists/receptors. Many of the interleukins (IL1A, IL2, IL4, IL10, IL12, IL17, IL18, IL23, sIL2R, sIL6R) were investigated in only a small number of studies (n = 1–5), and extracted data were therefore not suitable for meta-analysis. Of the interleukins, only IL1B and IL6 were measured in greater than five studies, with 10 and 23 studies, respectively, enabling meta-analysis. One study investigating SP IL1A concentration observed a significant decrease between control and azoospermic cohorts, but no difference was observed between the control and oligozoospermic cohorts ( Attia et al. , 2021 ). A second study reported a significant decrease in IL1A between controls and oligoasthenoteratozoospermia, but not between control and asthenoteratozoospermia or teratozoospermia ( Płaczkowska et al. , 2024 ). IL1B was investigated in 10 studies ( Fig. 5A ), with 3 studies reporting no difference between any infertile cohorts compared with controls ( Shimonovitz et al. , 1994 ; von Wolff et al. , 2007 ; Chyra-Jach et al. , 2018 ), 6 studies reporting a significant increase in IL1B in SP of infertile men ( Dousset et al. , 1997 ; Sanocka et al. , 2003 , 2004 ; Moretti et al. , 2023 ; Stabile et al. , 2023 ; Moretti et al. , 2025 ), and 1 study reporting a significant decrease in IL1B concentration in men with azoospermia compared with control men, but no difference in men with asthenozoospermia and oligoasthenoteratozoospermia ( Dziadecki et al. , 2010 ). Of these studies, only five were eligible for inclusion in the meta-analysis, which indicated no significant effect on the concentration of IL1B according to fertility status. The remaining five studies ( Dousset et al. , 1997 ; Sanocka et al. , 2003 ; Chyra-Jach et al. , 2018 ; Moretti et al. , 2023 , 2025 ) were not included in the meta-analysis as they reported SP IL1B concentrations as median (IQR). Of these five studies, one showed no change between the fertile and infertile groups ( Chyra-Jach et al. , 2018 ), while four reported an elevated concentration of IL1B in infertile men ( Dousset et al. , 1997 ; Sanocka et al. , 2003 ; Moretti et al. , 2023 , 2025 ). Bubble plot graphs visually depicting the concentrations of IL1B (A), IL6 (B), TNFA (C), and CXCL8 (IL8) (D) in seminal plasma, according to fertility status, in published studies included in the qualitative synthesis . Individual bubbles within panels indicate the component cohorts from each study, with bubble colour corresponding to the specific fertility status of the cohort as detailed in the legend. Each included study meets the eligibility criteria and has at least one fertile or normozoospermic control cohort and one infertile cohort. Bubble size corresponds to the number of participants in each cohort. Cytokine concentrations (mean or median values, according to format of published data) are shown in pg/ml, and individual studies are listed on the Y-axis in alphabetical order. Five studies investigated the abundance of IL2 in SP, with four studies reporting significantly higher IL2 concentration in infertile men compared to controls ( Paradisi et al. , 1995 ; Rajasekaran et al. , 1995 ; Matalliotakis et al. , 1998b , 2002 ), while one study reported no difference ( Dousset et al. , 1997 ). IL4 abundance was investigated in five studies ( Paradisi et al. , 1997 ; Omu et al. , 1999 ; Yiǧitbaşi et al. , 2010 ; Duan et al. , 2014 ; Płaczkowska et al. , 2024 ), with two studies reporting a significant reduction in IL4 concentration in infertile men compared with control men ( Duan et al. , 2014 ; Płaczkowska et al. , 2024 ), while the remaining three studies reported no differences ( Paradisi et al. , 1997 ; Omu et al. , 1999 ; Yiǧitbaşi et al. , 2010 ). The abundance of IL6 in SP was investigated in 23 studies ( Fig. 5B ). Of these, 15 formed the basis of the meta-analysis, which revealed a significant effect on the concentration of IL6 favouring an increase in infertile men compared with healthy controls. The remaining eight studies ( Dousset et al. , 1997 ; Sanocka et al. , 2003 ; Seshadri et al. , 2009 ; Castiglione et al. , 2014 ; Duan et al. , 2014 ; Chyra-Jach et al. , 2018 ; Tjagur et al. , 2021 ; Płaczkowska et al. , 2024 ) were not eligible to be included in the meta-analysis as they reported SP IL6 concentrations as median (IQR). Of the eight studies not included in the meta-analysis, three reported a significant increase in SP IL6 concentration in all infertile cohorts compared with healthy controls ( Sanocka et al. , 2003 ; Seshadri et al. , 2009 ; Castiglione et al. , 2014 ), while four reported no difference ( Dousset et al. , 1997 ; Chyra-Jach et al. , 2018 ; Tjagur et al. , 2021 ; Płaczkowska et al. , 2024 ), and one reported an elevated concentration of IL6, but only in the infertile cohort with accompanying genital tract inflammation ( Duan et al. , 2014 ). IL10 was investigated in five studies, two of which reported IL10 as being significantly reduced in SP from infertile cohorts, compared with fertile controls ( Camejo, 2003 ; Castiglione et al. , 2014 ), one study reported IL10 as being significantly reduced in men with oligoasthenoteratozoospermia compared to normozoospermic men, but not in men with teratozoospermia or asthenoteratozoospermia ( Firouzabadi et al. , 2024 ), one study reported no difference ( Rajasekaran et al. , 1996 ), and one study reported IL10 to be significantly increased in asthenozoospermic, oligoasthenozoospermic, and azoospermic cohorts compared with controls, but not in oligozoospermic men ( Seshadri et al. , 2009 ). IL11 was investigated in three studies ( Matalliotakis et al. , 1998c , 2002 ; Seshadri et al. , 2009 ). One study did not report on whether there was any difference in SP IL11 concentrations between the control and infertile cohorts ( Matalliotakis et al. , 1998c ). One study reported no difference between fertile and infertile men ( Seshadri et al. , 2009 ), and one study reported IL11 concentration as being significantly elevated only in SP of infertile men who also had evidence of a genital tract infection ( Matalliotakis et al. , 2002 ). IL12 was investigated in four studies, with three studies reporting IL12 as significantly reduced in SP of oligoasthenozsoospermic ( Chyra-Jach et al. , 2018 ), immunoinfertile ( Naz and Evans, 1998 ), and both general infertile and genital tract infection infertile ( Naz et al. , 1998 ) groups, whilst the fourth study reported no difference between cohorts ( Seshadri et al. , 2009 ). IL17 was investigated in two studies, with both reporting SP IL17 concentration as being elevated in infertile cohorts with accompanying diseases (either chronic genital tract inflammation or varicocele) compared with controls ( Duan et al. , 2014 ; Sabbaghi et al. , 2014 ). In addition, one of the studies reported IL17 as being elevated in the infertile cohort with no evidence of genital tract infection, compared with controls ( Duan et al. , 2014 ). IL18 was investigated in four studies, with all reporting no significant difference between infertile and control cohorts ( Matalliotakis et al. , 2006 ; Çalişkan et al. , 2010 ; Dziadecki et al. , 2010 ; Firouzabadi et al. , 2024 ). In one study however, a significant increase in IL18 SP was observed but only in the infertile cohort with evidence of genital infection ( Matalliotakis et al. , 2006 ). IL23 was investigated in two studies, with both reporting significantly increased IL23 concentration in the SP of infertile men, compared with controls ( Qian et al. , 2011 ; Duan et al. , 2014 ). sIL2R was investigated in four studies, three of which reported there being no difference in abundance between fertile and infertile men ( Dousset et al. , 1997 ; Paradisi et al. , 1997 ; Matalliotakis et al. , 1998b ), while one study described a significant increase in SP sIL2R concentration in men with asthenozoospermia, but not men with oligoasthenozoospermia ( Shimonovitz et al. , 1994 ). sIL6R was investigated in two studies, with one reporting no difference between the fertile and infertile cohorts ( Dousset et al. , 1997 ), while the second study reported significantly higher sIL6R abundance in infertile men with genital tract infection and idiopathic testicular lesions, compared with controls ( Matalliotakis et al. , 2000 ). Across the 52 studies, 6 chemokines were investigated, with 4 being members of the C-C ligand family, and 2 members of the C-X-C ligand family. All were investigated in relatively low numbers of studies, with CXCL8 being the most extensively investigated (14 studies spanning 1993 to 2024) ( Fig. 5D ). CCL2/MCP-1 abundance was investigated in two studies ( Shimoya et al. , 1995 ; Chyra-Jach et al. , 2018 ), with both reporting significantly higher SP concentrations in at least one of the infertile cohorts, compared with controls. CCL4/MIP-1B abundance in SP was investigated in one study, with the authors reporting a significant increase in concentration in the infertile cohorts compared with controls ( Chyra-Jach et al. , 2018 ). The abundance of CCL5/RANTES in SP was investigated by one study that found significantly reduced concentrations in a cohort of immune-infertile men (with anti-sperm antibodies in sera or semen) compared with controls, but no difference in any of the other infertile cohorts ( Naz and Leslie, 2000 ). One study investigated CCL11 abundance in SP, reporting significantly elevated concentrations in their infertile cohort compared with controls ( Yiǧitbaşi et al. , 2010 ). CXCL1/GRO was investigated in one study, which found no difference between fertile and infertile men unless the infertile cohort had accompanying leukocytospermia, which resulted in significant elevation of CXCL1/GRO ( Rajasekaran et al. , 1995 ). CXCL8 was investigated by 14 studies, 9 of which were included in a meta-analysis, which revealed a significant difference in the concentration of CXCL8 in SP between fertile and infertile cohorts ( Fig. 2C ). Five studies were deemed ineligible to be included in the meta-analysis due to CXCL8 abundance being reported as median (IQR) ( Sanocka et al. , 2003 ; Seshadri et al. , 2009 ; Chyra-Jach et al. , 2018 ; Lotti et al. , 2018 ; Płaczkowska et al. , 2024 ). Three of these studies reported CXCL8 as being significantly increased in SP of at least one of their infertile cohorts compared with controls ( Sanocka et al. , 2003 ; Seshadri et al. , 2009 ; Chyra-Jach et al. , 2018 ), whilst the remaining two studies reported no differences between control and infertile groups ( Lotti et al. , 2018 ; Płaczkowska et al. , 2024 ). Transforming growth factor beta 1 (TGFB1) was investigated in four studies, with one study quantifying only endogenously active (termed ‘bioactive’) TGFB1 ( Kisa et al. , 2008 ), two quantifying only total (acid activated) TGFB1 ( von Wolff et al. , 2007 ; Nilsson et al. , 2020 ) and one study quantifying both active and total TGFB1 ( Loras et al. , 1999 ). For TGFB1, three studies found no significant difference between control and infertile cohorts ( Loras et al. , 1999 ; von Wolff et al. , 2007 ; Nilsson et al. , 2020 ), while one study reported significantly increased bioactive TGFB1 in SP of men with oligoasthenoteratozoospermic, compared with controls ( Kisa et al. , 2008 ). The abundance of TGFB2 and TGFB3 in SP was investigated in only one study ( Nilsson et al. , 2020 ), which reported the concentration of total TGFB2 to be increased, and TGFB3 significantly decreased, in the infertile cohort compared with the control cohort. Growth and differentiation factor–15 (GDF-15) abundance in SP was investigated in one study, with the authors reporting no difference in concentration between fertile and infertile men ( Soucek et al. , 2010 ). Human Leukocyte Antigen G (HLA-G) was investigated in one study. No difference in HLA-G concentration was observed between infertile men and fertile controls ( Nilsson et al. , 2020 ). The abundance of colony-stimulating factor 1 (CSF1) (also known as macrophage colony-stimulating factor, M-CSF) in SP was investigated in one study that found no difference between control and male factor infertile cohorts, but a significant increase in immune-infertile men compared with controls ( Naz and Stanley, 1995 ). CSF2 (also known as granulocyte macrophage colony-stimulating factor, GM-CSF) was investigated in one study that reported significantly lower SP CSF2 concentration in a teratozoospermic cohort compared to controls, but no difference between control and asthenoteratozoospermic/oligoasthenoteratozoospermic cohorts ( Płaczkowska et al. , 2024 ). CSF3 (also known as granulocyte colony-stimulating factor, G-CSF) was investigated in one study that reported no difference in concentration between fertile and infertile men ( von Wolff et al. , 2007 ). Interferon gamma (IFNG) was investigated in five studies, three of which reported there being a significant increase in IFNG concentration in SP of infertile men ( Paradisi et al. , 1996 ; Duan et al. , 2014 ; Płaczkowska et al. , 2024 ), while two studies observed no differences ( Matalliotakis et al. , 2006 ; Seshadri et al. , 2009 ). The abundance of TNFA in SP was investigated in 13 studies ( Fig. 5C ), 9 of which fulfilled the criteria to form a meta-analysis. This analysis revealed a significant effect of fertility status on TNFA concentration, with infertile men having elevated SP TNFA compared with fertile counterparts ( Fig. 2B ). Four studies could not be included in the meta-analysis due to being reported as median (IQR) ( Sanocka et al. , 2003 ; Seshadri et al. , 2009 ; Castiglione et al. , 2014 ; Chyra-Jach et al. , 2018 ). Of the four remaining studies, one showed significant increases in infertile populations with prostatitis ( Castiglione et al. , 2014 ), whilst the other three studies found no differences between infertile and fertile cohorts ( Sanocka et al. , 2003 ; Seshadri et al. , 2009 ; Chyra-Jach et al. , 2018 ). One study largely reported TNFA as being undetected in SP as a median in both infertile and fertile men ( Seshadri et al. , 2009 ). The relationship between SP TRAIL concentration and fertility status was investigated in two studies, with both reporting TRAIL as being significantly increased in SP of infertile men compared with controls ( Duan et al. , 2014 ; Eid and Younan, 2015 ).

There is a large body of literature suggesting the potential utility of various cytokines as biomarkers for male fertility status. For many cytokines, there is insufficient data and/or too much heterogeneity between studies to draw meaningful conclusions on whether they are reproducibly informative. The current meta-analysis clearly identifies IL6, TNFA, and CXCL8 as consistently elevated in the SP of infertile men, compared with fertile men, particularly when an inflammatory condition exists. Further studies are required to confirm whether these cytokines are informative of fertility status or clinical conditions when other causes of infertility are excluded, and no cause of transient inflammation has been identified. The broader clinical relevance of these and other cytokines and chemokines will not be fully known until more high-quality studies are undertaken in well-described cohorts with appropriate participant numbers.

The imperative to identify biomarkers beyond bulk sperm parameters that are associated with and potentially informative of fertility and clinical status in men is an expanding area of clinical and research interest. There are more than 50 cytokines detectable in SP ( Lyons et al. , 2023 ), and given their association with disease states in other tissues, there is a strong prospect that altered concentrations or ratios of cytokines in SP are clinically relevant. As soluble immune-regulatory mediators, cytokines have established roles in the regulation of immune and inflammatory conditions of the reproductive tract, and many also regulate or have effects on sperm production, function, and developmental potential, and in eliciting and modulating a pro-tolerogenic female immune response to seminal fluid. While semen analysis will remain the cornerstone of male fertility evaluation, tests to evaluate cytokines in seminal fluid might have utility as part of clinical evaluation in couples experiencing infertility. However, their clinical utility and association or causal relationship to pathologies of the male reproductive tissues is not yet clear. In this systematic review and meta-analysis, we consolidated data from published studies comparing cytokine concentrations in SP of healthy fertile/normozoospermic men and men experiencing infertility. Data were extracted from a total of 52 eligible research studies published from 1993 to 2025, and a meta-analysis was performed to examine differences in IL1B, IL6, TNFA, and CXCL8 concentrations in SP of fertile and infertile men. A significant association was observed between fertility status and the concentrations of IL6, TNFA, and CXCL8, with infertile men demonstrating elevated levels of these cytokines, compared with fertile/normozoospermic controls. In contrast, IL1B did not show a significant effect in favour of either group, indicating no difference in SP IL1B concentration between fertile/normozoospermic men and infertile men. Across the 52 studies, a total of 30 individual cytokines and chemokines were reported as being detectable in SP of healthy fertile/normozoospermic and infertile men. Of these, only IL1B, IL6, TNFA, and CXCL8 satisfied the criteria to undergo individual random-effects meta-analysis to determine if their abundance differed between fertile/normozoospermic cohorts and infertile cohorts. Substantial disparity was observed between studies in the definitions employed to classify men as infertile. For example, some used the inability to conceive spontaneously despite having regular unprotected intercourse for greater than 12 months, while others were designated infertile based on sperm parameters falling below WHO reference ranges or decision limits. Others exhibited infertility associated with, and presumably caused by, inflammatory conditions and/or diseases that impact fertility, such as genital tract infection. When the infertile groups were combined irrespective of the underlying conditions, significantly elevated IL6, TNFA, and CXCL8 were observed in the SP of the infertile cohorts, compared with fertile/normozoospermic cohorts. IL6 is the most evaluated cytokine in SP to date, and the large number of high-quality studies on this cytokine adds confidence to the conclusion of its relationship to fertility status. Nevertheless, there was considerable heterogeneity in the IL6 meta-analysis, for which no obvious cause was evident. Analysis of the effects of small study bias, publication bias, and moderator effects in the full meta-analysis, wherein all infertile groups were combined, revealed no funnel plot asymmetry or clear bias. As the meta-analysis included studies spanning several countries across a large time period, some heterogeneity might be explained by differences in racial composition or population demographics. When studies were evaluated individually, 16 of 23 reported elevated IL6 concentration in SP from at least one infertile population compared with control cohorts, one study reported a significant decrease in IL6 concentration in SP of azoospermic males, and the remaining six studies reported no difference. For both IL6 and TNFA, sub-group analysis pointed to a stronger relationship with infertility when it was associated with an inflammatory condition, consistent with the known roles of these cytokines as pro-inflammatory agents ( Lampiao and Du Plessis, 2008 ; Fraczek et al. , 2013 ; Lotti and Maggi, 2013 ). This supports the interpretation that cytokines secreted from male reproductive tissue sites affected by microbial dysbiosis, infection, or sterile inflammation contribute to elevated SP cytokine levels. It also raises the question of whether undiagnosed or subclinical inflammatory conditions might contribute to the underlying mechanism when IL6 and/or TNFA are elevated in SP in men with unexplained infertility, but this speculation requires further investigation. A significant effect for CXCL8 was not seen when analysis was limited to studies where inflammatory disease was diagnosed, perhaps consistent with the prevalence of elevated CXCL8 in common as well as less frequently diagnosed reproductive tract conditions ( Lotti and Maggi, 2013 ). High levels of heterogeneity were again observed across the data sets in the subgroup meta-analysis, although outcomes were not changed when overly influential studies were removed. This implies that in some men, elevated inflammatory conditions are not causally related to infertility, while in other men, infertility is independent of inflammatory state. In many studies, other possible causes of infertility were not excluded. The relationship between elevated inflammatory cytokines and fertility—whether causal or co-factorial—would depend on the nature and duration of any underlying immune perturbation. It seems unlikely that short-lived infections or inflammatory responses could have lasting impacts on fertility, whereas chronic autoimmune or inflammatory conditions would reasonably be more likely to compromise immune tolerance of male gametes. It was not possible to specifically evaluate cytokines in idiopathic infertility in the current analysis, as there was extensive variation in the nature and analytical sensitivity of tests applied to exclude infection or inflammatory causes, and many older studies did not specify the tests conducted. In some men, inflammation may be present but clinically silent. Kopa et al. (2005) diagnosed a substantially higher proportion of neutrophil-associated inflammation in men undergoing fertility assessment by detecting SP elastase as opposed to peroxidase-positive cells (36% vs. 6.5%) and showed a correlation between SP elastase and IL6. For the meta-analysis, we applied strict eligibility criteria to include only studies with proven fertile and/or normozoospermic healthy controls as comparison groups. As a result, several studies that investigated the relationship between cytokines and chemokines in SP and inflammatory disease were excluded from the analysis due to the fertility status of control cohorts not being specified. Other studies point to a positive relationship between SP cytokines and inflammatory disease–particularly for cytokine CXCL8 ( Shimoya et al. , 1993 , 1995 ; Rajasekaran et al. , 1995 ; Depuydt et al. , 1996 ; Koumantakis et al. , 1998 ; Omu et al. , 1999 ; Sanocka et al. , 2003 , 2004 ; Moretti et al. , 2009 ; Dziadecki et al. , 2010 ; Kokab et al. , 2010 ; Lotti et al. , 2011a , 2011b ; Hajizadeh Maleki and Tartibian, 2021 ). Systematic reviews are a powerful tool for synthesizing existing evidence and informing decision-making, as well as identifying research knowledge gaps. A strength of this review is its systematic nature and the performance of meta-analyses. A pre-defined and replicable search strategy incorporating stringent eligibility criteria was employed to allow appropriate studies to be identified. These studies were then assessed by two independent reviewers to help reduce potential selection bias. Study quality was also assessed by two independent reviewers using the NHLBI-NIH quality assessment tool for observational cohorts and cross-sectional studies. Performing an objective meta-analysis allowed pooling of data from multiple studies to provide a more precise estimate of effect size. Additionally, for cytokines and chemokines where meta-analysis was able to be performed, rigorous testing for publication and small study bias, as well as individual study influence, was conducted to assess the reliability of the meta-analysis. A major limitation of this analysis is the small sample sizes in many included studies and high clinical and technical heterogeneity between studies. Many factors contribute to this heterogeneity, including the variable definitions of fertile/normozoospermic and infertility, variance in the underlying clinical conditions of participants, variance in likely confounding factors such as metabolic status and smoking, and changes over time to the WHO reference ranges/decision limits used to determine male fertility status. The substantial differences in clinical status of the infertile populations investigated, and considerable variation in the investigation of infertility causes, are key limitations. For example, Paradisi et al. (1997) reported 15 of 20 infertile men had varicocele, a higher proportion than in unselected groups of infertile men. All studies appeared independent other than the two Camejo et al studies (2001 and 2003), where the possibility of overlapping cohorts cannot be excluded. A sensitivity analysis showed that exclusion of these studies had no substantial bearing on the result (not shown). In future studies, it will be important to employ stringent criteria for defining fertile and infertile cohorts and subgroups, so that the relationship between elevated cytokines and specific conditions and their pathophysiological mechanisms can be delineated. We elected to include both proven fertile and healthy normozoospermic males with no prior history of infertility as control cohorts. Although pooling control groups may contribute to heterogeneity, the approach was considered an acceptable compromise to maximise data inclusion. When the specific status of the control cohort was included as a moderator effect in the meta-analyses, no effect was found. This is consistent with our previous study focussed on SP cytokines in healthy men, where no effect of including both proven fertile and normozoospermic men was observed ( Lyons et al. , 2023 ). It is important to note that because the control groups contain normozoospermic men without proven fertility, the prospect that some of these men are functionally infertile despite normal sperm parameters cannot be excluded. Indeed, given that normal sperm parameters do not guarantee fertility, and the pathophysiology of conditions that give rise to this circumstance is not clear, there is a prospect that subsets of normozoospermic men with reduced fertility have altered SP cytokine profiles. If this is the case, it does not compromise the conclusions drawn herein, as any confounding effect would reduce, rather than increase, the effect size between healthy control and infertile groups. It will be of great interest to explore this prospect in future studies, as biomarkers of reduced fertility that are independent of sperm parameters would have high clinical utility. Another major contributor to the heterogeneity observed between studies is the lack of standardised methodology for processing and storage of SP, and in the assay platform for cytokine quantification ( Lyons et al. , 2023 ). The development and adoption of an agreed set of guidelines to enable a standardised approach for preparing and storing SP and quantifying cytokine abundance is now a priority. Another important limitation relates to data extraction methodology. In some cases, we were unable to extract individual participant data, so data from the whole cohort, descriptive statistics, or graphical representations thereof were utilized. For some studies, we chose not to derive the mean and standard deviation from median, range, and IQR descriptive statistics. While it would have been preferable to include all possible datasets, the inclusion of median, range, and IQR can lead to skewing of the data. Finally, we performed statistical analysis to evaluate the influence of biases within the meta-analysis and found no evidence for this. However, biases such as publication bias are inherently complex, being comprised of a variety of factors that can lead to bias that cannot truly be accounted for, meaning the potential presence of publication bias cannot be definitively excluded. Further, since all effect measures were unadjusted for confounders, this may influence the validity of pooled estimates and limit causal inference. A considerable research effort has focused on determining the identity of cytokines and chemokines in the SP of men and quantifying their abundance to examine whether differences exist between fertile and infertile men. Despite this, the understanding of whether and how seminal fluid cytokines directly contribute to male fertility and successful conception is limited. An important issue identified in this systematic review is the high level of heterogeneity that exists between studies. This is a key contributing factor to the often-conflicting findings reported in different studies, and a barrier to understanding the significance of SP cytokines and their potential clinical utility. To progress the field, there is an imperative to determine the identity and biological function of those factors that are associated with and potentially informative of male fertility status. To achieve this, high-quality, appropriately powered studies using thorough clinical investigation and standardised methodologies for the collection, processing, and storage of SP samples are required. A consensus in the field as to the most accurate, sensitive, specific, and reproducible assay platform for SP cytokine and chemokine quantification is needed. Reasonably, such a platform must also offer high throughput and sufficient ease of use to be suitable for clinical implementation. Once these conditions are met, it will be possible to better understand the precise relationship between SP cytokines and the pathophysiology of male infertility, and to consider whether cytokines might have utility in diagnostic tests to improve clinical semen analysis. A key unresolved question is whether cytokines might have prognostic or diagnostic utility in men where no infection or inflammatory condition is identified, but infertility is unexplained. IL6, TNFA, CXCL8, and/or other cytokines could be sensitive indicators of subclinical immune and inflammatory conditions affecting sperm function or developmental competence. To determine this, will require carefully selected cohorts where other causes of infertility and overt transient causes of elevated inflammatory cytokines are excluded. Another consideration is to understand the tissue origins of SP cytokines and their individual relationships to specific conditions in the male reproductive tract and systemically. Cytokines in SP might originate from the peripheral blood after crossing the blood-testis barrier, as well as arising from local secretion by the prostate, seminal vesicles, bulbourethral gland, and epididymis, and somatic cells of the testis ( Hedger and Meinhardt, 2003 ; Huleihel and Lunenfeld, 2004 ; Drabovich et al. , 2014 ; Fraczek and Kurpisz, 2015 ). The correlation between cytokine levels in blood serum and SP has been examined, but findings are often inconsistent ( Havrylyuk et al. , 2015 ; Bongrani et al. , 2019 ; Płaczkowska et al. , 2024 ). Most recently, Placzkowska et al. , reported that different sperm disorders (i.e. terato-, asthenoterato-, and oligoasthenoteratozoospermia) are linked to distinct cytokine profiles in SP and serum, highlighting the potential for the influence of both local and systemic factors on semen parameters. Importantly, IL6 and CXCL8 were found to be orders of magnitude higher in SP than serum, even when the authors excluded participants with inflammation attributable to infection and/or leukocytospermia ( Płaczkowska et al. , 2024 ), although asymptomatic inflammation could not be excluded. Previous studies have linked elevated SP CXCL8 to prostatitis, particularly in men with high BMI ( Lotti et al. , 2011a , 2011b ). Immune and inflammatory disturbances associated with infection, metabolic status, age, diet, and micronutrient deficiencies, hormone imbalance, loss of immune tolerance, and structural abnormalities in male reproductive tissues might contribute to SP cytokine profile ( Fraczek and Kurpisz, 2015 ; Loveland et al. , 2017 ; Ma et al. , 2025 ). Future research should therefore focus on uncovering the local and systemic tissue sources and mechanisms shaping SP cytokine composition as well as defining their specific roles in the pathophysiology of male infertility. Well-controlled, high-quality studies with clearly defined clinical groups are now needed to clarify whether specific patterns of cytokines can be linked with different clinical conditions.