New Research in Male Fertility

Three things, and only one of them is a treatment. First, the reference standard: the WHO laboratory manual's 6th edition was published on 27 July 2021 and moved away from fixed reference thresholds, which changes how a semen report should be read. Second, the evidence on sperm-selection add-ons consolidated — Cochrane's June 2026 update found that of seven strategies, one has moderate-certainty evidence for reducing miscarriage and none can yet be recommended for routine use. Third, a randomised trial in 1,064 couples showed that ICSI does not beat conventional IVF when the man's count and motility are normal, which is an argument for doing less rather than more.

Almost everything else described as a breakthrough in this field is at the stage of laboratory work, single small trials, or reviews calling for trials. That is not a criticism of the research; it is the difference between a finding and a treatment, and it is the distinction this page is organised around.

The test to apply to any new claim is simple and it is the one the guideline bodies apply: has this been randomised, against what comparison, and did it change live birth or only a semen number? For what is currently offered in clinics, the treatment-by-treatment picture runs alongside this page.

The numbers moved slightly; the bigger change is conceptual. The 6th edition's lower fifth percentiles, derived from men whose partners conceived within a year of trying, are: semen volume 1.4 mL (95% CI 1.3 to 1.5), total sperm number 39 million per ejaculate (95% CI 35 to 40), total motility 42% (95% CI 40 to 43), progressive motility 30% (95% CI 29 to 31), vitality 54% (95% CI 50 to 56) and normal forms 4% (95% CI 3 to 4). The equivalent 5th edition values from 2010 were 1.5 mL, 39 million, 40% total motility, 32% progressive motility, 58% vitality and 4% normal forms (Boitrelle et al, Life 2021;11(12):1368).

The conceptual change matters more. The 6th edition abandons the notion of reference thresholds and suggests replacing them with decision limits — but, as the critical review notes, proposes no decision limits for basic, extended or advanced parameters. So a result below the fifth percentile is not a diagnosis and a result above it is not a clearance; both are positions in a distribution of men who conceived. One practical change: vitality testing is now advised when total motility is below 40%, where the 5th edition tied it to progressive motility.

This is the single most useful thing to know before reading a report, because the older 2010 numbers still circulate widely and the two sets get mixed. What a semen analysis does and does not establish covers how the parameters relate to each other.

Mostly not measurably, and the best-studied one reduces miscarriage rather than raising births. Cochrane's 2026 update searched to 24 February 2025 and included five randomised trials with 3,752 women, having removed four previously included studies after applying a trustworthiness assessment to their data. For hyaluronic-acid-selected ICSI against standard ICSI: live birth per woman randomised RR 1.09 (95% CI 0.97 to 1.24), described as no difference or a slight increase, on low certainty evidence from one trial of 2,772 participants; miscarriage per woman randomised RR 0.59 (95% CI 0.44 to 0.80) from two trials and 3,327 participants, on moderate certainty evidence; and clinical pregnancy RR 0.98 (95% CI 0.90 to 1.07), little or no difference (Garg et al, Cochrane Database of Systematic Reviews 2026, CD010461).

Microfluidic sorting, zeta-potential selection and SpermSlow each rest on a single small trial at very low certainty — for microfluidic sorting, live birth RR 1.10 (95% CI 0.68 to 1.76) from 122 participants. The review's conclusion is that the effect of these other techniques on live birth, miscarriage and clinical pregnancy is unknown, and that further high-quality studies, including data awaited from 19 ongoing trials, are needed before any can be recommended for routine practice.

The underlying large trial is worth naming because it is the one that settled the live birth question. HABSelect randomised 2,772 couples at 16 UK assisted conception units to hyaluronan-selected ICSI or standard ICSI; term live birth was 27.4% (379 of 1,381) versus 25.2% (346 of 1,371), OR 1.12 (95% CI 0.95 to 1.34, p=0.18), and the authors state that wider use is therefore not recommended (Miller et al, Lancet 2019;393:416-22). The HFEA, whose add-on ratings were last reviewed on 26 February 2026, rates PICSI black, meaning the evidence indicates no effect on treatment outcome, and rates IMSI grey for insufficient evidence.

No, on the only large randomised test of the question. In an open-label trial at two centres in Ho Chi Minh City, 1,064 couples in whom the male partner had normal total sperm count and progressive motility were randomised to ICSI or conventional IVF. Live birth after the first embryo transfer from the initiated cycle occurred for 184 of 532 couples (35%) with ICSI and 166 of 532 (31%) with conventional IVF — an absolute difference of 3.4% (95% CI -2.4 to 9.2), RR 1.11 (95% CI 0.93 to 1.32), p=0.27. Fertilisation failure was 5% with ICSI and 6% with IVF. The authors write that their results challenge the value of routine ICSI in this population (Dang et al, Lancet 2021;397:1554-63).

ICSI use has grown worldwide largely in couples without male factor infertility, which is what makes this trial consequential: it is evidence for not adding a step. Where ICSI is indicated is a separate question, covered in what ICSI is for, and the wider add-on question in which IVF add-ons have evidence.

Surgical retrieval works for some men and the figures have not been improved by anything newer. A systematic review of seven comparative studies found overall sperm retrieval ranged from 42.9% to 63% with microdissection TESE versus 16.7% to 45% with conventional TESE in non-obstructive azoospermia, with the advantage clearest in patchy patterns such as Sertoli cell-only, less favourable in uniform maturation arrest, and fewer complications on ultrasound after microTESE. The same review records that randomised trials comparing the two are lacking (Deruyver et al, Andrology 2014;2(1):20-4). A later meta-analysis puts retrieval by microTESE at 40% to 60% (Tharakan et al, Human Reproduction Update 2022).

Hormonal priming before retrieval is the most common addition and it does not hold up in general. Pooling 22 studies and 1,706 participants, hormone therapy before surgical sperm retrieval was associated with a higher retrieval rate overall (OR 1.96, 95% CI 1.08 to 3.56), but the subgroup analysis found the improvement only in normogonadotropic men (OR 2.13, 95% CI 1.10 to 4.14) and not in hypergonadotropic men (OR 1.73, 95% CI 0.44 to 6.77). The literature was at moderate or severe risk of bias, and the authors conclude hormone therapy should not be used routinely before retrieval, with large randomised trials still needed.

The genetic work-up, by contrast, is settled enough to be specified. The EAU's 2026 guideline recommends karyotype analysis and genetic counselling for all men with azoospermia or sperm concentration below 5 million/mL (Strong), Y-chromosome microdeletion testing at concentrations of 1 million/mL or below with consideration below 5 million/mL (Strong), and CFTR testing for men with structural abnormalities of the vas deferens and their partners (Strong). The AUA/ASRM guideline recommends Y-chromosome microdeletion analysis for primary infertility with azoospermia or concentration at or below 1 million/mL (Moderate Recommendation, Grade B). These are the genetic tests that exist in practice, as opposed to the panels sometimes described as personalised medicine.

It probably helps pregnancy rates and the live birth evidence is genuinely uncertain. Cochrane's 2021 review included 48 studies with 5,384 participants. Against delayed or no treatment, treatment may improve pregnancy (RR 1.55, 95% CI 1.06 to 2.26; 13 RCTs, 1,193 participants, low certainty) — against a 21% chance with no or delayed treatment, between 22% and 48% after surgical or radiological treatment. For live birth, only two trials with 204 participants reported it and they pointed in different directions, leaving RR 2.27 (95% CI 0.19 to 26.93) at very low certainty (Persad et al, Cochrane Database of Systematic Reviews 2021, CD000479).

On technique, microscopic subinguinal repair probably improves pregnancy slightly compared with other surgical approaches (RR 1.18, 95% CI 1.02 to 1.36; 12 RCTs, 1,473 participants, moderate certainty) — the clearest comparative finding in the review. Surgical versus radiological treatment remains uncertain for both pregnancy and adverse events.

Both guidelines restrict the indication rather than widening it. The EAU 2026 guideline recommends treating infertile men with a clinical varicocele, abnormal semen parameters and otherwise unexplained infertility where the female partner has good ovarian reserve (Strong). The AUA/ASRM guideline recommends repair for men with palpable varicoceles, infertility and abnormal semen parameters (Moderate Recommendation, Grade B) and states clinicians should not recommend varicocelectomy for non-palpable varicoceles (Strong Recommendation, Grade C).

The two major guidelines disagree, and knowing that is more useful than a verdict. The AUA/ASRM guideline on diagnosis and treatment of infertility in men, published in 2020 and amended in 2024, states at statement 16 that clinicians should not recommend sperm DNA fragmentation analysis in the initial evaluation (Moderate Recommendation, Evidence Level Grade C), while recommending it in recurrent pregnancy loss (Moderate Recommendation, Grade C). The EAU's 2026 guideline recommends performing DNA fragmentation testing in couples with recurrent pregnancy loss from natural conception, failure of assisted reproduction, or unexplained male infertility (Strong).

So the honest position is that the test has a place in specific situations that both bodies recognise — recurrent loss and treatment failure — and no established place as a first test for a couple who have just started investigating. The EAU separately advises against routine reactive oxygen species testing in diagnosis and management (Weak).

It is also worth knowing what a high result would change. The EAU lists varicocelectomy as something that may be considered in men with raised DNA fragmentation and otherwise unexplained infertility or ART failure, but rates that recommendation Weak. What DNA fragmentation testing measures covers the assays themselves.

The largest analysis says yes and is contested on method. A meta-regression combining 223 studies and 288 estimates from semen samples collected between 1973 and 2018 reported that, in the pooled estimate, mean sperm concentration among unselected men fell 51.6% over that period (-1.17 million/mL/year, 95% CI -1.66 to -0.68), with total sperm count down 62.3%. It was the first such analysis to report a decline among unselected men from South and Central America, Asia and Africa (-0.65 million/mL/year, 95% CI -1.29 to -0.01, p=0.045), and the post-2000 slope was steeper, with the percentage decline per year rising from 1.16% to 2.64% (Levine et al, Human Reproduction Update 2023;29(2):157-176).

A published critique argues that this literature rests on a set of implicit assumptions it calls the Sperm Count Decline hypothesis, identifies weaknesses and inconsistencies in it, and proposes an alternative framework in which sperm count varies across a wide, largely non-pathological range, so that population averages cannot be read directly as a measure of fertility or health (Boulicault et al, Human Fertility 2022;25(5):888-902).

For an individual the practical consequence is small either way: a population trend does not describe one man's result, and the 2021 manual's fifth percentiles are the reference for reading his. The trend matters for research priorities and public health, which is where the argument is being had.

The following are research, and describing them otherwise is where pages like this usually go wrong.

• Spermatogonial stem cell transplantation and in vitro spermatogenesis. No human clinical trial was retrievable for this article; searches of the NCBI databases for first-in-human or clinical trials of spermatogonial stem cell transplantation returned no records, and what exists is laboratory culture work and reviews of it. The previous version of this page stated that early results from clinical trials were promising; that claim has been removed.
• Artificial intelligence in semen analysis and sperm selection. What was retrievable is review articles describing potential, not validation studies or randomised trials reporting live birth.
• Microfluidic sperm sorting. One randomised trial of 122 participants, live birth RR 1.10 (95% CI 0.68 to 1.76), rated very low certainty by Cochrane in 2026.
• Gene editing. No clinical application in male infertility exists, and the reference to it has been removed rather than hedged.
• Antioxidant and nutraceutical regimens as a treatment. Cochrane's 2022 review of 90 studies and 10,303 subfertile men rates live birth evidence very low certainty and calls the overall evidence inconclusive; the AUA/ASRM guideline calls the benefits of supplements of questionable clinical utility (Moderate Recommendation, Grade B).
• Proteomic and metabolomic sperm profiling. Research-stage with no guideline recommendation retrieved.

None of this means the work is worthless. It means a man being offered any of it today is being offered research, and is entitled to be told so. Where supplements specifically are concerned, what the trials on male fertility supplements measured sets out the outcomes they actually reported.

Against the claims commonly made for new male fertility research, the retrieved sources do not support the following.

• That advanced sperm selection produces more live births. Cochrane 2026 finds low-certainty no-difference for the best-studied method and unknown effects for the rest, and the 2,772-couple HABSelect trial found term live birth 27.4% versus 25.2%, p=0.18.
• That ICSI is the safer default. In 1,064 randomised couples with normal count and motility, live birth was 35% with ICSI versus 31% with conventional IVF, RR 1.11 (95% CI 0.93 to 1.32).
• That hormone therapy before sperm retrieval helps generally. The pooled benefit across 22 studies and 1,706 participants was confined to normogonadotropic men, at moderate to severe risk of bias, and the authors advise against routine use.
• That DNA fragmentation testing belongs in an initial work-up. AUA/ASRM recommends against it there; EAU recommends it in recurrent loss, ART failure and unexplained male infertility. Both restrict it to defined situations.
• That a semen result above the WHO fifth percentile means fertility, or below it means infertility. The 6th edition moved away from thresholds and proposed no decision limits to replace them.
• That falling population sperm counts predict an individual's chance of conceiving. The decline estimate is contested on method, and neither side claims it applies to one man's result.
• That stem cell treatment, gene editing or AI selection is available. No retrieved human trial supports clinical use of any of them.

What is established is narrower and more useful: across the comparative series and the later pooled estimate, microdissection TESE retrieves sperm in roughly 40% to 63% of men with non-obstructive azoospermia; microscopic subinguinal varicocelectomy probably gives slightly better pregnancy rates than other surgical approaches; karyotype, Y-chromosome microdeletion and CFTR testing have guideline-grade indications; hyaluronic-acid sperm selection probably reduces miscarriage without changing live birth; and ICSI adds nothing when the sperm are normal.