Stress and IVF: What the Evidence Clears You Of — and What Is Actually Trainable

Evidence-based · Written by Dr. Lejla Fazlicic, D.Ac, L.Ac · All key claims cited to peer-reviewed research

What remains is a narrower and more useful question: chronic physiological load, which is measurable and, unlike worry, genuinely trainable. Slow-paced breathing raises vagally-mediated heart rate variability in a way that accrues with practice across 223 studies. Regular physical activity lowers cortisol with moderate-certainty evidence. Protecting sleep restores the recovery half of the stress cycle. What no study has ever demonstrated is that moving any of those markers changes a birth. So the practices are worth doing because they make three and a half months of waiting more survivable, and because the axis is demonstrably suppressible by sustained load — not because anyone has shown they buy you a baby.

Working out which of these fits your actual life, rather than an idealised one, is the hard part to do alone.

Book a free 15-minute call · See what working together looks like

"Just relax" is not wrong. It is incomplete. And it is usually delivered by people who love you and have nothing else to offer.

Did stress cause my IVF to fail?

No. And this is one of the few places in reproductive medicine where the evidence is clean enough to say so without hedging.

In 2011, a meta-analysis in the BMJ pooled 14 prospective psychosocial studies covering 3,583 infertile women, each undergoing a single ART cycle. The pooled standardised mean difference in pre-treatment anxiety and depression between women who conceived and women who did not was −0.04, with a confidence interval from −0.11 to 0.03 — and every subgroup analysis was null (Boivin et al., 2011).

Two later meta-analyses are sometimes offered as contradicting it. Read closely, they do not test the same claim:

• One found small associations with clinical pregnancy — but those associations were not significant for serum pregnancy tests or live birth rates, and the fail-safe number did not exceed criterion in any analysis, meaning every result is fragile to publication bias (Matthiesen et al., 2011).
• Another found associations for distress measured during treatment rather than before it — and found nothing at all for change in distress, with zero heterogeneity (Purewal et al., 2018). Distress during a cycle is downstream of how the cycle is visibly going. A poor response and a cancelled transfer both make people anxious.

If nobody told you this before your last cycle, that is not your failure. Nobody tells anyone. And no habit, no worry and no unmanaged week ever decided an outcome on its own — a point we take apart properly in was the failed cycle my fault?

Do stress-reduction programs improve IVF outcomes?

This is where a lot of expensive things get sold, so it is worth being exact.

The Cochrane review of psychological and educational interventions for subfertile men and women examined 39 studies and 4,925 participants undergoing ART — and the authors declined to pool the data at all, judging that a meta-analysis would not produce a clinically meaningful estimate. Only 2 of the 39 studies reported live birth, both with substantial attrition, and the combined odds ratios ranged from 1.13 to 10.05 — an interval so wide it tells you nothing. Their verdict was that trials were "generally poorly designed and executed" (Verkuijlen et al., 2016).

You will see a much more exciting number quoted: a relative risk of 2.01 for clinical pregnancy from a 39-study meta-analysis (Frederiksen et al., 2015). Take that seriously for a moment. It would mean counselling roughly doubles your odds of conception. Nothing in reproductive medicine does that — not a change of protocol, not a change of clinic, not PGT-A. When an intervention appears to outperform every actual treatment, the most likely explanation is the quality of the trials feeding the estimate, and Cochrane examined largely the same trials and reached the opposite conclusion.

An earlier meta-analysis found a pregnancy-rate benefit that applied only to couples not receiving ART, alongside no measurable effect on mental health — a combination with no plausible mechanism (Hämmerli et al., 2009).

Even the best-known mind-body trial in IVF tells the story if you read it in order. Among 143 women randomized before their first cycle, only 9% of the intervention group had attended half their sessions by cycle 1 — and cycle 1 pregnancy rates were identical at 43%. The widely quoted 52%-versus-20% figure comes from cycle 2, among the smaller, self-selected group who failed cycle 1 and stayed in the study (Domar et al., 2011). The properly randomized, fully-exposed comparison was null.

What these programs do reliably do is reduce distress. Cochrane's median standardised mean difference for depression was −0.45. That is a real benefit, and it is enough on its own. It is simply not a pregnancy claim.

And acupuncture? I am a licensed acupuncturist, so let me be direct.

The evidence does not support acupuncture as a way to improve IVF outcomes. The largest trial randomized 848 women across 16 IVF centres to acupuncture or sham and found live births of 18.3% versus 17.8% (Smith et al., 2018). A meta-analysis restricted to trials using a non-invasive placebo device pooled 1,835 women and found a live birth relative risk of 0.87 with a confidence interval of 0.75 to 1.01 — high-certainty evidence, zero heterogeneity, and no significant difference in either direction (Coyle et al., 2021). Cochrane reached the same conclusion (Cheong et al., 2013).

People frequently find acupuncture calming during a brutal process, and that is a different claim from a live birth claim. I do not make the second one. I go through this in more detail — including where the trials could fairly be criticised — in what nobody can tell you yet about preparing for IVF.

Does cortisol affect egg quality?

Here is where the picture becomes more interesting, and more useful.

The cleanest demonstration comes from an experimental study measuring GnRH directly in pituitary portal blood in follicular-phase ewes. An acute six-hour cortisol elevation did nothing to GnRH pulse frequency. A sustained 27-hour elevation reduced GnRH pulse frequency by 45% and delayed the preovulatory LH surge by 10 hours — and by 70% in an artificial follicular phase (Oakley et al., 2009).

Duration, not peak. That single finding reframes the whole conversation. A hard afternoon is not the variable. Months without recovery might be. The neuroanatomy connecting stress signals to GnRH neurons is real but, as the reviewers themselves note, not yet fully constructed (McCosh et al., 2022).

At the extreme end of that spectrum sits functional hypothalamic amenorrhea — chronic anovulation associated with stress, weight loss, excessive exercise, or some combination, recognised in an Endocrine Society clinical practice guideline co-sponsored by ASRM (Gordon et al., 2017). It establishes that the axis can be shut down by sustained load. It does not establish that sub-clinical stress in an ovulating IVF patient does anything measurable, and it would be dishonest to slide from one to the other.

What the biomarker studies found — including the nulls

This is the part that keeps the section honest.

What the stress biomarker studies actually found, including the null results

Marker	Study	Result
Salivary cortisol, women trying to conceive	LIFE cohort, 373 couples	No association with fecundability
Salivary alpha-amylase, women trying to conceive	LIFE cohort, 373 couples	Highest tertile: 29% reduction in fecundity (ratio 0.71, upper bound touching 1.00)
Both markers, pregnancy loss	Same LIFE cohort	No clear pattern of association
Salivary cortisol, IVF patients	135 women, four separate indices	All four null — awakening response p=0.485, AUC-ground p=0.527, AUC-increase p=0.731, diurnal slope p=0.889
Hair cortisol (3–6 month load), IVF patients	88 women, one clinic	Predicted clinical pregnancy, p=0.017, ~27% of variance
Hair cortisol, attempted replication	43 women, pilot	Coefficient ran in the opposite direction
Hair cortisol, testicular function	696 men	No association — and no association with self-reported stress either

The hair cortisol result is the only one that even gestures at a link between chronic load and IVF outcome (Massey et al., 2016). It rests on 88 women from roughly 800 eligible at a single clinic, the endpoint was clinical pregnancy rather than live birth, independent experts called it likely a chance finding at the time of publication, the one attempted replication went the other way (Santa-Cruz et al., 2020), and the same lead author's earlier systematic review found three studies showing higher cortisol was more favourable and five showing lower cortisol was (Massey et al., 2014).

That is the dashed link, and it is worth drawing rather than hiding. Salivary cortisol — the marker sold in home test kits — is null in every relevant study. Alpha-amylase has one positive natural-conception finding and one hypothesis-generating IVF study (Zhou et al., 2019). Nobody should be interpreting their own stress biomarkers, and no one should be selling you a way to.

What about him?

Stress research on the male side follows the same shape: the experience is real, the reproductive consequence is unproven.

In 1,215 young men from the general population, self-reported stress above an intermediate level was associated with poorer semen quality in a dose-response manner — 38% lower sperm concentration in the highest versus intermediate stress group. Crucially, there were no significant associations with any reproductive hormone (Nordkap et al., 2016). A cohort of 193 men found perceived stress inversely associated with sperm concentration and morphology, and two or more stressful life events in a year associated with a lower motile fraction — while job strain was associated with nothing (Janevic et al., 2014). And in 286 men attending an infertility clinic, high and medium occupational stress were both associated with a higher DNA fragmentation index, though the paper reports p-values without effect sizes (Radwan et al., 2016).

Then the same Danish group measured hair cortisol in 696 men and found no association with testicular function, and no association between hair cortisol and self-reported stress either. Their own conclusion is that any negative reproductive effects of stress "may not be mediated directly by cortisol" (Nordkap et al., 2022).

And in 423 men, those from infertile couples reported significantly more stressful life events than men whose partners had conceived naturally — 50.4% versus 36.9% — but stress was not associated with testicular function in any of them (Bräuner et al., 2020). Infertility is stressful. That is not the same as stress being the cause. No blame in either direction, in either partner.

What actually lowers cortisol before IVF?

Recovery is a skill — and skills can be built. Here is what has randomized evidence behind it, stated at exactly the level the evidence supports. Every one of these stops at a physiological marker or a distress score. None has been shown to change a birth.

Recovery practices with randomized evidence — and the marker each one has been shown to move

Practice	Best evidence	What actually moves
Slow-paced breathing	223 studies pooled	Vagally-mediated heart rate variability rises during the session, immediately after a single session, and after a multi-session program — the training effect is the point
HRV biofeedback	24 studies, 484 participants	Self-reported stress and anxiety, Hedges g = 0.83 versus control. A distress result, not a biomarker one
Meditation / mindfulness	45 RCTs with active controls	Cortisol, C-reactive protein, blood pressure, heart rate, triglycerides. Focused-attention practice specifically reduced cortisol
— calibrated by	Meta-analysis of RCTs measuring cortisol change	Cortisol reduction was significant only in at-risk samples; in 21 saliva studies of unselected people it was small and non-significant
Yoga (asana-based)	42 RCTs with active controls	Evening and waking cortisol, blood pressure, resting heart rate, fasting glucose. Note high-frequency HRV went down, contrary to the usual framing
Physical activity	10 intervention studies	Cortisol SMD −0.37, moderate certainty; sleep quality SMD −0.30, low certainty
Protecting sleep	Controlled sleep-loss experiment	One night of partial sleep loss raised next-evening cortisol 37%; total loss 45%; onset of the axis's quiet period delayed by at least an hour

Sources in order: Laborde et al., 2022; Goessl et al., 2017; Pascoe et al., 2017; Koncz et al., 2021; Pascoe et al., 2017 (yoga); De Nys et al., 2022; Leproult et al., 1997.

One nuance worth keeping. The cortisol meta-analysis restricted to randomized trials found the effect concentrated in people whose load was genuinely elevated, and small to absent in unselected samples. If you are three months out from a cycle costing $15–35,000, you are plausibly in the first group. That is the supportive reading — and it is also why nobody should promise the general population a cortisol drop.

What the evidence does not show

• That your distress before a cycle affects its outcome. Fourteen prospective studies, 3,583 women, effect size −0.04.
• That psychological interventions raise pregnancy rates. Cochrane reviewed 39 studies and declined to pool; only two reported live birth.
• That acupuncture improves live birth. High-certainty evidence, 1,835 women, relative risk 0.87 with a confidence interval of 0.75 to 1.01 versus placebo needling — no significant difference in either direction.
• That cortisol is the mechanism. Salivary cortisol is null in every relevant study; hair cortisol in 696 men showed no association with testicular function.
• That moving HRV or cortisol changes a fertility outcome. Not one study in this literature demonstrates it. Every trainability finding stops at the biomarker, and this article stops there too.
• That there is a home test worth buying. There is not.

What we do with this

Modify the modifiable. Measure the measurable. On stress, that produces a shorter and more honest list than most sites offer:

1. Put the blame down first. This is not a soft opening — it is the best-evidenced statement in the article, and carrying guilt into a cycle costs you something real while buying nothing.
2. Choose duration over intensity. The physiology points at sustained load, not at bad days. Ten minutes of slow breathing most days beats one heroic weekend retreat, and the breathing evidence specifically shows the effect accruing with repetition.
3. Use movement as your primary lever. It has the cleanest certainty rating of anything here — moderate-certainty cortisol reduction plus a sleep benefit — and it acts on several other systems at once. We map those in exercise in the 90 days before IVF.
4. Protect sleep as a stress intervention, not just a sleep one. Sleep loss degrades the recovery half of the cycle specifically, delaying the onset of the axis's quiet period by at least an hour. More on the evidence in sleep before IVF.
5. Take support because it helps you, not because it helps the embryo. Psychological support reliably reduces distress. That is a sufficient reason. If you are struggling, that belongs with a professional, and there is nothing about IVF that makes it a weakness to ask.
6. Do the practices with him. His stress data is no stronger than yours, which is exactly the point — neither of you is the problem, and both of you are in the same three and a half months.

Before the window: live your life. Inside it: do your part. After transfer: it is out of your hands.

And the sentence that matters most on the morning of the cycle is not "I stayed calm." It is: I did my part. The rest is not in my hands.

Related reading

• Was the failed IVF cycle my fault?
• Why did my IVF fail if everything looked normal?
• Sleep before IVF: what the evidence actually shows
• Exercise in the 90 days before IVF
• What nobody can tell you yet about preparing for IVF
• The research library — stress section

Frequently asked questions

Does stress cause IVF failure?

No. A meta-analysis of 14 prospective studies covering 3,583 women found that pre-treatment emotional distress did not predict whether an IVF cycle succeeded — the pooled standardised mean difference was −0.04, with a confidence interval crossing zero, and every subgroup analysis was null. Later meta-analyses that found small associations measured distress during treatment rather than before it, where reverse causation is the obvious explanation, and found nothing for change in distress.

Do stress-reduction or mind-body programs improve IVF success rates?

They reliably reduce distress, and that benefit is real. They have not been shown to improve pregnancy or live birth. The Cochrane review of 39 studies and 4,925 ART participants declined to pool the data, found only two studies reporting live birth, and judged the trials generally poorly designed. The best-known IVF mind-body trial found identical pregnancy rates of 43% in both arms at the cycle where randomisation was intact.

Does stress during the two-week wait affect implantation?

No study has shown that it does. The distress research that found associations measured distress during treatment rather than before it, and reverse causation is the obvious explanation — a cycle that is visibly going badly makes people anxious. The one meta-analysis that examined change in distress from baseline through treatment found no association with outcome, with zero heterogeneity between studies. The two-week wait is genuinely hard, and there is no evidence that finding it hard changes what happens.

Can lowering cortisol improve egg quality?

No study has shown this. Salivary cortisol showed no association with fecundability in a preconception cohort of 373 couples, and all four salivary cortisol indices were null in 135 women undergoing IVF. One study of hair cortisol in 88 women found an association with clinical pregnancy, but the single attempted replication ran in the opposite direction and the same author's earlier systematic review found studies pointing both ways. Cortisol home testing is not interpretable.

What actually lowers stress physiologically before IVF?

Three things have randomized evidence. Slow-paced breathing raises vagally-mediated heart rate variability, and pooled across 223 studies the effect accrues with repeated practice rather than being momentary. Regular physical activity lowers cortisol with a standardised mean difference of −0.37, graded moderate certainty. Protecting sleep matters specifically because sleep loss raises the following evening's cortisol by 37 to 45% and delays recovery of the stress axis. None of these has been shown to change a fertility outcome.

Does stress affect sperm quality?

The associations exist but the mechanism does not hold up. In 1,215 men, self-reported stress above an intermediate level was associated with a 38% lower sperm concentration, and in 286 infertility-clinic patients occupational stress was associated with higher DNA fragmentation. But the same research group measured hair cortisol in 696 men and found no association with testicular function at all — concluding that any effects of stress may not be mediated directly by cortisol. In 423 men, those from infertile couples reported more stressful life events, yet stress was not associated with testicular function in any of them.

---

This article is educational and reflects associations reported in the scientific literature. It is not medical advice, not a diagnosis, and not a way to interpret your own results. As a Doctor of Acupuncture I offer lifestyle guidance that works alongside your physician and fertility clinic — I do not diagnose, interpret labs, or prescribe. If you are struggling emotionally, that is common in fertility treatment and worth raising with your clinic or a mental health professional.