Coffee is the least harmful of the three, alcohol shows a measurable dose-response even at one drink a day, and smoking is the only one that damages egg quality, sperm quality, implantation and pregnancy maintenance at once. The pooled IVF evidence is reassuring for caffeine — a live-birth odds ratio of 0.98 across 26,922 women and couples — while the miscarriage evidence at high intake is not.
This article separates four questions people merge into one: does the substance change natural conception, does it change IVF outcomes, does it change whether a pregnancy survives, and does the male partner’s exposure matter. Figures come from meta-analyses and prospective cohorts indexed in PubMed, plus patient guidance from ACOG and the WHO semen reference manual. Written by the ProIVF Medical Editorial Team and reviewed by the ProIVF Medical Advisory Board; ProIVF sells no products and is affiliated with no clinic named here.
What Does Each Substance Actually Cost You?
Caffeine barely moves IVF success at normal intakes; smoking moves almost every reproductive endpoint at once.
| Exposure | What the pooled data show | Direction |
|---|---|---|
| Caffeine, usual IVF intakes | Live birth OR 0.98 (0.89–1.08); pregnancy OR 0.97 (0.85–1.12) | Essentially null |
| Caffeine 300 mg/day | Miscarriage RR 1.37 (1.19–1.57) | Harm signal |
| Caffeine 600 mg/day | Miscarriage RR 2.32 (1.62–3.31) | Strong harm signal |
| Any alcohol vs none | Natural fecundability RR 0.87 (0.78–0.95) | Harm |
| Each extra 12.5 g/day of alcohol | Fecundability RR 0.98 (0.97–0.99) | Linear harm |
| 84 g/week alcohol | Women OR 0.93 (0.90–0.98); men, measured in the partner OR 0.91 (0.88–0.94) | Both partners |
| Smoking (men) | Count −9.72 million/mL, motility −3.48%, morphology −1.37% | Harm |
| Smoking (women, in ART) | Live birth per cycle OR 0.54 (0.30–0.99); miscarriage OR 2.65 (1.33–5.30) | Harm |
Start with the exposure carrying the largest evidence-backed penalty and the lowest cost to change — tobacco — and treat alcohol as a dose question. Our IVF preparation checklist sets out what to change by cycle week.
Does Coffee Stop You Getting Pregnant?
In IVF specifically, no meaningful effect has been shown: the largest pooled analysis found caffeine intake was associated with a live-birth odds ratio of 0.98, statistically indistinguishable from no effect.
What the IVF meta-analysis found
A 2022 Acta Obstetricia et Gynecologica Scandinavica review of 12 caffeine and 14 alcohol studies pooled 26,922 women and couples undergoing IVF or ICSI. For caffeine:
- Pregnancy OR 0.97 (95% CI 0.85–1.12) when the female partner was exposed, 0.93 (0.75–1.14) when the male partner was.
- Live-birth OR 0.98 (0.89–1.08) for women and 0.98 (0.86–1.12) for men.
- Every interval crossed 1.0, so none of these estimates demonstrates harm.
That is the honest headline for anyone asking whether a morning coffee is costing them a cycle: in these numbers, it is not.
Why the miscarriage data say something different
A separate meta-analysis of caffeine and spontaneous abortion found a relative risk of 1.37 (95% CI 1.19–1.57) at 300 mg of caffeine per day and 2.32 (1.62–3.31) at 600 mg, with no clear association with time-to-pregnancy.
The same authors add a caution that the IVF numbers alone would never surface:
“The advised limit of a maximum of two to three cups of coffee/200–300 mg caffeine per day may be too high.” — Clinical Epidemiology, 2017 meta-analysis
Read together, the two literatures point one way: caffeine does not appear to prevent implantation, but very high intake is associated with losing a pregnancy after it starts. That argues for a cap, not for giving up coffee.
How much caffeine is in what you actually drink?
A brewed filter coffee holds roughly 80–120 mg, instant 60–80 mg, a café large brew can exceed 150 mg, black tea 40–70 mg, green tea 20–45 mg, a can of cola 30–40 mg and an energy drink 80–160 mg. Two café lattes and a cola is therefore about 300 mg — already the level where the miscarriage association appears.
ACOG and most fertility clinics use 200 mg/day as the working ceiling; WHO and EFSA figures are higher, and the 2017 meta-analysis above is one reason several authors argue that ceiling is too generous for someone actively trying to conceive.
Does Alcohol Lower IVF Success Rates?
The IVF evidence is ambiguous, but the natural-conception evidence is not: across 19 studies and 98,657 women, any alcohol intake was associated with 13% lower fecundability versus abstinence.
Three studies, three different answers
| Study | Population | Finding |
|---|---|---|
| Scientific Reports 2017 | 98,657 women, natural conception | Fecundability RR 0.87 any drinking; 0.89 light; 0.77 moderate-heavy |
| AOGS 2022 meta-analysis | 26,922 women/couples in IVF/ICSI | Maternal alcohol pregnancy OR 0.83 (0.69–1.01); paternal alcohol partner live birth OR 0.88 (0.79–0.99) |
| Human Reproduction 2019, Danish cohort | 1,708 women; 1,511 IUI, 2,870 IVF/ICSI, 1,355 frozen cycles | Live-birth aRR 1.00 (0.83–1.21) at 1–2 drinks/week; 0.95 (0.75–1.20) at 3–7; 0.89 (0.53–1.51) above 7 |
The Danish study is the one most often quoted to argue that a glass of wine is harmless, and its IVF/ICSI figures genuinely show no step-down across intake bands. What it did not test is conception without treatment, where the same body of evidence does show a gradient.
What counts as one drink, and where the dose-response starts
A standard drink is about 14 g of alcohol: 355 mL of 5% beer, 150 mL of 12% wine, or 45 mL of spirits. In the pooled natural-conception analysis risk rose linearly — relative fecundability 0.98 per additional 12.5 g per day, roughly a 2% loss of per-cycle chance per extra drink, with no threshold where the slope flattened.
Drink type mattered far less than amount (wine RR 0.98, beer 1.02, spirits 0.92) — worth remembering when a clinic suggests switching from beer to wine rather than drinking less.
Does the male partner’s drinking matter?
Yes, and it is the finding most patients have not heard. In the 2022 IVF meta-analysis, paternal alcohol was associated with the female partner’s live birth at OR 0.88 (0.79–0.99), and at roughly 84 g per week the odds ratios were 0.93 (0.90–0.98) for women and 0.91 (0.88–0.94) for men.
Because semen quality recovers over roughly one spermatogenic cycle, a reduction taken three months before a cycle is more defensible than one taken the week before it.
No safe level of alcohol in a pregnancy has been established, and IVF is a process that can end in one, so the weeks around transfer and the two-week wait are treated differently from the rest. Our IVF diet guide covers what changes between stimulation and transfer.
How Much Does Smoking Damage Fertility?
Smoking is the only lifestyle exposure in this article that harms eggs, sperm, implantation and pregnancy maintenance simultaneously — and the only one where stopping has a documented recovery timeline.
What it does to sperm
A meta-analysis of 20 studies in 5,865 men found smokers’ sperm concentration 9.72 million/mL lower (95% CI 13.32–6.12 lower), motility 3.48 points lower (5.53–1.44) and normal morphology 1.37 points lower (2.63–0.11). Against the WHO 2021 limits of 16 million/mL and 39 million total, that can move a man across the line — and about 37% of men of reproductive age smoke.
Our semen analysis guide explains how to read a report.
What it does to eggs, AMH and menopause
A 2024 review of 65 studies from 15,072 records reported current smokers’ AMH 0.4 to 1.1 ng/mL below non-smokers, a relative difference of 4% to 44%. The authors’ caveat belongs in the same sentence: AMH is a marker, and how far a smoking-related shift reflects true ovarian reserve is unresolved.
The endpoint data are harder to argue with. In InterLACE — 17 prospective studies across seven countries, 207,231 women, current smokers had twice the risk of premature menopause (RRR 2.05, 95% CI 1.73–2.44) and 80% higher risk of early menopause (1.80, 1.66–1.95); former smokers sat in between (1.13 and 1.15), and at 11–15 pack-years the figures were 4.35 (2.78–5.92) and 3.01 (2.15–4.21).
What it does to an IVF cycle and a pregnancy
An older but still widely cited meta-analysis of 21 studies put smokers’ live birth per ART cycle at OR 0.54 (0.30–0.99), clinical pregnancy at OR 0.56 (0.43–0.73), miscarriage at OR 2.65 (1.33–5.30) and ectopic pregnancy at OR 15.69 (2.87–85.76). The ectopic interval is very wide and the review predates modern vitrification and frozen transfer, but no later meta-analysis has reversed the direction.
So a clinic asking about smoking is asking about all four pathways at once, not egg quality alone; the same logic runs through how to improve egg quality and our male infertility and IVF guide.
Are Vaping and Cannabis Safer Choices?
Neither has reassuring data, and cannabis has a specific disclosure problem: in one Canadian study, 59% of patients with cannabis detected in follicular fluid had not reported using it at intake.
What the e-cigarette data actually show
In a Danish study of 2,008 young men (median age 19.0 years), 52% smoked cigarettes, 13% used e-cigarettes, 25% used nicotine snuff and 33% used marijuana. Daily e-cigarette users had higher total sperm counts than non-users (147 vs 91 million), and daily cigarette smokers did too (139 vs 103 million).
Those comparisons are cross-sectional, unadjusted for abstinence, and internally inconsistent — which is the point. The vaping literature neither supports e-cigarettes as protective nor condemns them; it is thin. Nicotine itself carries the reproducible sperm and vascular effects, whatever the device.
What cannabis does, and what it hides
A study of 318 follicular fluid samples from 261 IVF patients in Toronto found 17 (6.4%) cannabis-positive, with positivity rising from about 4% to 12% after Canadian legalisation in October 2018. Ten of those 17 — 59% — had not disclosed cannabis use when asked at intake, and the same study measured changes in cannabinoid receptor expression and global DNA methylation.
Outcomes in users are less clear-cut. In a clinic series of 722 IVF patients, 68 (9.4%) used cannabis, mostly lightly: implantation 40.74% versus 41.13% and ongoing pregnancy 35.2% versus 29.1% — numerically better in users, and almost certainly about who uses and why. A Journal of Urology review concluded cannabis may harm male fertility, on evidence its authors called insufficient.
The practical instruction is short: tell your clinic, and stop before the cycle. Non-disclosure is the documented failure mode here, not mild use.
How Long Before a Cycle Do I Need to Stop?
Three months is the window that matters most for sperm, and the current cycle is the window that matters for alcohol and caffeine.
| Exposure | Minimum lead time | Why |
|---|---|---|
| Cigarettes and nicotine | 3 months, ideally 6 | Spermatogenesis takes about 74 days plus transit; oocyte recruitment runs to months |
| Alcohol | 3 months for him, whole cycle for her | Semen parameters recover per cycle; no safe level exists in a pregnancy |
| Caffeine | From the start of stimulation | The outcome at stake is loss after transfer |
| Cannabis | 1–3 months | Alters follicular environment and sperm parameters; disclosure matters meanwhile |
| Secondhand smoke | Immediately | Dose-dependent and avoidable at zero cost |
Nobody should read this table as a reason to postpone an indicated cycle. If you are 38 with declining reserve, three months of perfect behaviour costs more fertility than it buys: quit what you can this week — see IVF success rate by age.
If Quitting Works, Why Do Lifestyle Trials Fail?
Because in randomised trials, lifestyle advice did not improve live birth: a 2021 Cochrane review of 7 randomised trials in 2,130 participants found a live-birth relative risk of 0.93 (95% CI 0.79–1.10) with preconception counselling.
This is the most misread result in reproductive lifestyle research. It does not mean smoking and drinking are harmless; it means the intervention tested — advice, counselling, written plans — did not move the outcome in patients already treated for infertility and already motivated.
Three consequences follow for how you should use the evidence in this article:
- Observational effects are smaller than they look: most figures above are odds ratios between 0.87 and 0.98, and confounding by why a person drinks or smokes is never fully removed.
- The credible signals are the steep ones: 600 mg of caffeine (RR 2.32) and 11–15 pack-years (RRR 4.35) far outweigh a 12% shift at moderate intake.
- Advice is not change: pharmacological and behavioural support outperform information alone.
Lifestyle is where you start, not where the plan ends. The few nutrients with randomised data are in our fertility supplements guide, and the one lifestyle variable with its own trial evidence — and its own surprises — is exercise during IVF.
Three Patients, Three Different Habits
The following patient stories are shared with consent. Names and identifying details have been changed to protect privacy.
Case 1: Rui, 35 — five caffeinated drinks a day, two cycles (Zhengzhou)
Rui drank two Americanos each morning, a milk tea after lunch and a cola most evenings, and never added it up; her first stimulation cycle cost ¥36,000 and ended in a clinical pregnancy lost at 7 weeks.
Her difficulty was that nobody had asked her a quantity. “Every leaflet said less coffee and I heard don’t drink coffee, so I kept drinking it and felt guilty about it,” she said.
Counting turned four drinks into roughly 520 mg of caffeine a day, and her clinic asked her to come under 200 mg before the next transfer. Her second cycle cost ¥34,500, she transferred at 36, and her son was born at 38 weeks. “The change that mattered was boring: I swapped two of the coffees for decaf and stopped the cola.”
Case 2: Tomas, 39 — swapped cigarettes for a vape (Valencia)
Tomas had smoked 15 cigarettes a day for 16 years and switched to an e-cigarette when the clinic told him to stop, assuming that settled it; his first semen analysis showed 9 million/mL against the WHO line of 16 million.
His difficulty was that the substitution was invisible to the people treating him. “I told them I’d quit smoking and I meant it. Nobody asked what I was smoking instead,” he said.
He stopped nicotine entirely 11 weeks before retrieval, at about €260 a month in replacements and lost productivity he would have spent on cigarettes anyway. His second analysis was 17 million/mL, and one daughter followed after two ICSI cycles at €6,200 each. “Counting my sperm as a two-year project was the only version of this that worked.”
Case 3: Sana, 31 — cannabis for nausea, undisclosed (Toronto)
Sana used cannabis twice weekly for endometriosis pain and nausea, paid roughly CAD$13,000 for her IVF cycle, and answered “no” when the intake form asked about recreational drugs.
Her difficulty was shame, not the substance: the clinic’s own screening found cannabinoid exposure in her follicular fluid after retrieval, and her doctor raised it in a follow-up call rather than before the cycle.
She disclosed fully at the next consultation, stopped before the second retrieval, and transferred once at 32 — a boy born at 39 weeks. “One sentence on a form changed nothing about my cycle. Not saying it changed what my doctor could advise.”
FAQ
Q: How much coffee is safe during IVF?
Most clinics use 200 mg per day — about one and a half to two café brewed cups — because the miscarriage association appears at 300 mg (RR 1.37, 95% CI 1.19–1.57) and strengthens sharply at 600 mg.
Q: Can I drink alcohol during an IVF stimulation cycle?
The pooled IVF data are ambiguous — one large Danish cohort found a live-birth rate ratio of 1.00 (0.83–1.21) at one to two drinks a week — but natural-conception data show 13% lower fecundability with any intake, and no safe level once pregnant.
Q: Is vaping better than smoking for fertility?
There is no evidence that it is: among 2,008 young men, daily e-cigarette users had higher total sperm counts than non-users (147 vs 91 million), an inconsistent cross-sectional finding that supports neither reassurance nor condemnation.
Q: Does cannabis stop IVF treatment?
It should not, and disclosure is what protects you: in a Toronto study of 318 follicular fluid samples, 6.4% were cannabis-positive and 10 of those 17 patients (59%) had not disclosed use at intake.
Q: Will my clinic judge me for smoking or drinking?
Clinics need the quantity, not the confession: every threshold in this article is dose-based — 200 mg of caffeine, 14 g of alcohol per drink, and 3 months for a sperm sample to recover.
Planning Your Next Step
Rank the exposures by evidence rather than guilt: tobacco first, alcohol as a dose question, caffeine above roughly 200 mg, and cannabis mainly as a disclosure issue. Three months is the window where sperm change; the current cycle is where pregnancy-loss risk lives.
If you want a clinic that takes the male partner’s history seriously, filter our reproductive centre directory by country and services, or contact our team for a shortlist. Evidence here moves quickly and circumstances move faster; your treating team outranks any published average.
Written by the ProIVF Medical Editorial Team and reviewed by the ProIVF Medical Advisory Board. Evidence base: meta-analyses and systematic reviews indexed in PubMed in Acta Obstetricia et Gynecologica Scandinavica, Human Reproduction, Human Reproduction Update, European Urology, Clinical Epidemiology, Scientific Reports, PLOS Medicine, the Journal of Urology, the Journal of Cannabis Research and Cochrane Database of Systematic Reviews, plus the InterLACE pooled analysis and patient guidance from ACOG and the WHO semen laboratory manual. ProIVF sells no supplements or cessation products and has no commercial relationship with any clinic or brand named here.
Last updated: 25 September 2026. This article is educational content and is not a substitute for individualised medical advice. Always follow the specific guidance of your fertility clinic and treating physicians.
