How Much Coffee Can I Actually Drink Before IVF?

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

The honest caveat sits elsewhere: the miscarriage literature is less settled than the IVF literature, and it is a separate question. Two other things will be more useful to you than any number. First, cups are not a unit — the same order from the same coffee shop on six consecutive days has been measured at anywhere from 259 to 564 mg. Second, if you are going to give up one drink, the evidence says give up the sugary one, not the coffee.

Working out where you actually sit on that scale is harder than it sounds.

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

If you quit coffee before your last cycle and it did not work, and you have quietly wondered whether you slipped on a Tuesday — that is not what decided it. The numbers below were available the whole time; almost nobody hands them to a patient, and that is a failure of the handout, not of you.

Does caffeine affect IVF success rates?

The best available answer is a 2022 systematic review and dose-response meta-analysis covering 26,922 women and their spouses undergoing IVF or ICSI. Every caffeine estimate was null (Rao et al., 2022):

• Women's caffeine → pregnancy: odds ratio 0.97 (0.85–1.12)
• Women's caffeine → live birth: 0.98 (0.89–1.08)
• Men's caffeine → pregnancy: 0.93 (0.75–1.14)
• Men's caffeine → live birth: 0.98 (0.86–1.12)

Their conclusion, verbatim: "There was no association between caffeine consumption and pregnancy or live birth rate of IVF/ICSI."

Here is what makes that null result persuasive rather than merely reassuring. The same paper, using the same methods on the same pool of studies, found alcohol was not null. Maternal alcohol and pregnancy came in at 0.83 (0.69–1.01) — itself crossing the line — but paternal alcohol and live birth reached 0.88 (0.79–0.99), and the dose-response analysis found a 7% lower pregnancy chance in women and a 9% lower live-birth chance in men at 84 g of alcohol a week. The method could detect a signal. It detected one for alcohol and none for caffeine.

Two large single cohorts agree. Among 2,474 couples across 4,716 IVF cycles, live birth odds ratios were 1.00, 0.89 and 1.07 across rising caffeine categories, with a trend p-value of 0.74 (Choi et al., 2011). And in the Harvard EARTH cohort, adjusted live birth per initiated cycle ran 46%, 44%, 42%, 40% and 40% across caffeine categories from under 50 mg/day to over 300 mg/day — trend p-value 0.34 (Abadia et al., 2017).

Where the "quit coffee" advice actually comes from

Almost every clinic instruction traces back to one 2002 study of 221 couples. It reported roughly three- to fourfold higher odds of not achieving live birth with female caffeine intake — but read the exposure definition. There were two exposed categories, more than 2 to 50 mg a day and more than 50 mg a day. The lower one is a square of chocolate; the upper one is less than a single cup of coffee. The reference group was 0–2 mg a day, meaning near-total abstainers, and the confidence intervals ran from 1.1 to 11.6. The authors themselves wrote: "If these findings are replicated, caffeine use should be minimized prior to and while undergoing IVF/GIFT" (Klonoff-Cohen et al., 2002).

They were not replicated. Four larger attempts — 2,474 couples, 300 women across 493 cycles, 340 women, and a 26,922-person meta-analysis — all failed to reproduce it.

What do the guidelines actually say?

This is where the gap between the evidence and the advice becomes visible.

What each body actually says about caffeine and fertility — note that none of them says "quit"

Body	Verbatim position
NICE NG257 (March 2026)	"Inform people who are concerned about their fertility that there is no consistent evidence of an association between consumption of caffeinated beverages (tea, coffee, energy drinks and colas) and fertility problems."
ACOG Committee Opinion 462, reaffirmed 2023	"Moderate caffeine consumption (less than 200 mg per day) does not appear to be a major contributing factor in miscarriage or preterm birth… A final conclusion cannot be made at this time as to whether there is a correlation between high caffeine intake and miscarriage."
ASRM Optimizing natural fertility, 2022	"Moderate caffeine consumption (1–2 cups of coffee per day or its equivalent) before or during pregnancy has no apparent adverse effects on fertility or pregnancy outcomes. Caffeine consumption has no effect on semen parameters in men." (ASRM's concern threshold is 500 mg/day, above five cups.)
EFSA 2015 safety opinion	"Habitual caffeine consumption up to 400 mg per day does not give rise to safety concerns for non-pregnant adults." "Habitual caffeine consumption up to 200 mg per day by pregnant women does not give rise to safety concerns for the fetus."

Notice two things. In the NICE guideline, alcohol and smoking both receive restrictive recommendations and caffeine receives a reassurance — and that reassurance was reviewed and kept in the 2026 edition. And ACOG has never told anyone to quit: it says below 200 mg there is no apparent major effect, and above it "a final conclusion cannot be made." Somewhere between the guideline and the handout, we do not know became don't.

How much caffeine is actually in my coffee?

This is the part that changes what you do, because most people are not where they think they are.

Approximate caffeine per serving — and note how wide the real-world ranges are

Drink	Typical	Measured range
Brewed filter coffee, 8 oz / 240 mL	~95 mg	FDA gives 113–247 mg per 12 oz
Coffee-shop brewed, 16 oz	~188 mg average	259–564 mg measured from the same outlet on six consecutive days
Single espresso, 30 mL	~64–106 mg	25–214 mg across 97 Australian samples; a separate 20-sample UK analysis found a 6-fold spread between outlets, with the highest single shot at 322 mg
Instant coffee	No official per-serving figure	Varies substantially with scoop size and brand
Black tea, 8 oz	~48 mg	FDA: 71 mg per 12 oz
Green tea, 8 oz	~29 mg	FDA: 37 mg per 12 oz
Matcha	Higher than brewed green tea	No official per-serving figure exists — it is whole powdered leaf, so treat it closer to coffee than to tea
Cola, 12 oz	FDA range only	23–83 mg
Energy drink, 12 oz	FDA range only	41–246 mg
Decaf coffee, 8 oz	~2 mg	FDA: 2–15 mg

Three measurement studies make the same point independently. Twenty high-street espressos in Glasgow analysed by HPLC showed a six-fold difference between outlets, with the highest single shot at 322 mg (Crozier et al., 2012). Ninety-seven Australian espresso samples ranged from 25 to 214 mg per serve, with 12% above 167 mg (Desbrow et al., 2007). And US specialty coffees ranged 58 to 259 mg per dose — with the same beverage from the same outlet on six consecutive days measuring 259 to 564 mg (McCusker et al., 2003).

So "two cups" is not a dose. Two cups can be 120 mg or it can be over 600 mg. If you want to stay under a threshold, the threshold has to be in milligrams, and the practical move is to know your usual drink rather than to count mugs.

What about caffeine and miscarriage?

This is the part where the reassurance thins, and you deserve it straight rather than smoothed.

Three separate meta-analyses find a dose-response association between caffeine in pregnancy and pregnancy loss. The most rigorous — restricted to prospective studies, 14 studies, 130,456 participants — found relative risks of 1.02 at 50–149 mg/day, 1.16 at 150–349, 1.40 at 350–699 and 1.72 at 700 or more (Chen et al., 2016). Read carefully: the two lowest categories, up to 349 mg/day, both cross the null. Significance appears only from about 350 mg. That is three to four strong cups. The authors' own conclusion opens with the word "albeit inconclusive."

The reverse-causation problem, measured rather than assumed

Two studies published in the same year asked the same question and got opposite answers, and the difference between them is the most instructive thing in this literature.

One found caffeine at 200 mg or more associated with roughly double the miscarriage hazard, and stated that nausea did not materially affect the association (Weng et al., 2008). That paper drew four separate published letters of challenge in the same journal.

The other assessed caffeine at three time points with explicit attention to how intake changed. Across 2,407 pregnancies and 258 losses, coffee and caffeine at all three points were unrelated to miscarriage risk — and caffeine looked harmful only among women whose loss had already happened before they reported their intake (Savitz et al., 2008).

The mechanism is simple and human. A viable pregnancy produces high hCG, which produces nausea, which produces coffee aversion. A non-viable pregnancy produces less of all three. So surviving pregnancies look like low-caffeine pregnancies whether or not caffeine did anything at all.

Two further lines of evidence support that reading. A 2022 meta-analysis found case-control studies produced effect estimates roughly 1.5 to 2 times larger than cohort studies for the identical exposure — which is recall bias, quantified. And a Mendelian randomisation study using genetic instruments for coffee consumption across 49,996 miscarriage cases found "no change in risk of sporadic miscarriages, stillbirths, pre-term birth" (Brito Nunes et al., 2023).

That is not a debunking, and I will not call it one. Mendelian randomisation for coffee uses weak genetic instruments and a noisy self-reported outcome. The honest statement is that the association is real in the observational data, it is contested, and the best design available for testing causation does not reproduce it.

Does caffeine affect egg quality?

Something measurable does happen, and it is worth saying so rather than claiming caffeine is inert.

Caffeine physically reaches the follicle. In 619 women undergoing IVF with a mean intake of 456 mg/day, caffeine was measurable in both serum and follicular fluid — and after adjustment, there was still no association with pregnancy success rate (Al-Saleh et al., 2010).

The strongest challenge to the null comes from a 2026 study that used biomarkers rather than questionnaires: 737 women, 2,211 urine samples across three repeated measures, plus 680 follicular fluid samples. Higher urinary caffeine was associated with about 4.85% fewer oocytes retrieved and 6.78% fewer mature oocytes (Luo et al., 2026). Biomarkers beat questionnaires, so this deserves weight.

Two things temper it. The endpoints are intermediate — no clinical pregnancy or live birth outcome was reported, and that is precisely where the 26,922-person meta-analysis found nothing. And the biomarker approach does not consistently show harm: a study of 1,228 women that assayed serum caffeine, paraxanthine and theobromine found no association for any of the three. Comparing top to bottom tertile, fecundability odds ratios were 0.87 for caffeine, 0.92 for paraxanthine and 1.15 for theobromine, all crossing the null. The authors concluded that "caffeine exposure from usual low to moderate caffeinated beverage intake likely does not influence fecundability" (Purdue-Smithe et al., 2022).

Does my partner need to give up coffee?

ASRM states flatly that "caffeine consumption has no effect on semen parameters in men," and the largest data agree: men's caffeine and live birth, odds ratio 0.98. In 2,554 young Danish men, caffeine intake up to 800 mg/day showed no association with semen quality (Jensen et al., 2010).

Two honest counterpoints. In 80 healthy men, those drinking more than three cups a day had about 20% more double-strand sperm DNA damage on a neutral comet assay — a laboratory marker, in a study designed to examine age, with total abstainers as the comparison group (Schmid et al., 2007). And a 2025 EARTH analysis found lower adjusted live birth in IVF cycles for men in the top tertile of coffee and tea — but the same analysis found men drinking the most beer had the highest live birth, which is a strong signal that these beverage-specific findings reflect lifestyle rather than the beverage.

The systematic review of 28 papers and 19,967 men concludes the male literature is "inconsistent and inconclusive" (Ricci et al., 2017). That is the right summary. The rest of his list — what actually is worth changing on the male side — is in what does he actually have to give up before IVF?

Are soda and energy drinks worse than coffee for IVF?

Four independent datasets split coffee from soft drinks, and in each of them the signal lands on the soft-drink side rather than the coffee side.

• In 340 women undergoing IVF, sugared soda drinkers had fewer oocytes, fewer mature oocytes and fewer top-quality embryos, with live birth 12–16 percentage points lower — while coffee, caffeine and diet soda showed nothing (Machtinger et al., 2017).
• In 3,628 Danish women trying to conceive, soda showed falling fecundability ratios of 0.89 down to 0.48 across servings, while coffee at three or more servings a day was 1.05 (Hatch et al., 2012).
• In 2,554 Danish men, heavy cola drinkers had sperm concentrations of 40 million/mL versus 56 in non-cola drinkers — while consuming under 140 mg of caffeine a day, so caffeine cannot be the mechanism.
• Ricci's systematic review of 28 papers and 19,967 men found coffee, tea and cocoa did not affect semen parameters while cola and caffeinated soft drinks did (Ricci et al., 2017).

And the honest counterweight, because it is the largest and most recent study to look at these beverages together. A prospective cohort of 612 women across 1,572 treatment cycles found that intake of both sugar-sweetened and artificially sweetened beverages was "unrelated to the likelihood of success in infertility treatment" — adjusted live birth of 0.41 versus 0.41 across the lowest and highest intake categories in IVF and ICSI cycles (Salas-Huetos et al., 2026).

So: if one drink is coming out of your day, the sugary one is the better candidate on the balance of the evidence. That is a cheap and reasonable swap, not a settled finding.

Should I quit cold turkey when my cycle starts?

No — and this is a cost worth counting. A critical review of 66 studies found caffeine withdrawal produces headache in 50% of people and clinically significant distress or functional impairment in 13%, with onset 12 to 24 hours after stopping, peak intensity at 20 to 51 hours, and duration of 2 to 9 days. Symptoms appear from doses as low as 100 mg/day, and expectancy effects are not the explanation (Juliano & Griffiths, 2004).

That window lands squarely on stimulation, monitoring and retrieval. No study has tested whether withdrawal affects cycle outcomes, and I am not going to claim it does. But it is a certain, quantified cost during the hardest fortnight of the process, incurred for a benefit the IVF literature does not support. If you want to cut down, taper in the weeks beforehand rather than on the morning of day one.

What the evidence does not show

• That caffeine is inert. It reaches follicular fluid; a biomarker study found a small oocyte-yield reduction; a small study found a sperm DNA marker. The claim is not that nothing happens — it is that nothing has been shown to change whether you take home a baby.
• That the miscarriage association has been debunked. Three meta-analyses find it. Mendelian randomisation does not reproduce it. Contested is the accurate word.
• That quitting caffeine improves anything. I could not find a trial that has tested a caffeine intervention against a fertility endpoint; every study here is observational.
• That "two cups" means anything. The same order has been measured at 259 and 564 mg on consecutive days.
• That decaf is caffeine-free. It is 2 to 15 mg per cup — small, not zero.

What we do with this

1. Under about 200 mg a day, no guideline sets a limit and no replicated study supports concern. ACOG, EFSA, ASRM and NICE all converge here — and the one study that ever did find a signal at that level has not been reproduced in four larger attempts.
2. Between 200 and 350 mg is personal-choice territory. The best prospective miscarriage meta-analysis finds nothing below 350 mg, and the IVF literature finds nothing at all.
3. Above about 350 mg is where I would draw a line — not because of the IVF data, which is null, but because that is where the miscarriage meta-analyses start to show something and the downside of moving is close to zero.
4. Know your actual drink. A home filter coffee is about 95 mg. A large coffee-shop brew averages 188 mg and has been measured far higher. That single fact moves more people than any threshold.
5. Cut the sugary drink before the coffee. Four datasets point that way; the largest and most recent found nothing for either. It is the better candidate, not a proven one.
6. If you are cutting down, taper. Not on the morning of stimulation.
7. Your partner is under essentially no restriction from any guideline. ASRM says so explicitly.

Related reading

• What do you actually have to give up before IVF?
• Which pregnancy rules apply before transfer?
• Alcohol before IVF: how long should you stop?
• What to eat before IVF: a 90-day guide
• Egg quality in the 90 days before retrieval
• The research library

Frequently asked questions

How much coffee can I drink before IVF?

Below roughly 200 mg of caffeine a day — about one large brewed coffee or two smaller ones — no guideline sets a limit and no replicated study supports concern. ACOG, EFSA, ASRM and NICE all converge there. Between 200 and 350 mg the IVF evidence remains null and the best prospective miscarriage meta-analysis shows nothing. Above about 350 mg is where the miscarriage meta-analyses begin to show a signal, and a reasonable place to draw a line.

Does caffeine affect IVF success rates?

The largest analysis found no effect. A 2022 dose-response meta-analysis of 26,922 IVF and ICSI patients found caffeine associated with neither pregnancy (odds ratio 0.97) nor live birth (0.98) in women, nor in men (0.93 and 0.98). The same paper, using identical methods, did find an alcohol association — showing the method could detect a signal. Two large single cohorts of 2,474 couples and 300 women reached the same null conclusion.

Does caffeine cause miscarriage?

This is contested rather than settled. Three meta-analyses find a dose-response association, but the best prospective one shows nothing below about 350 mg a day. Reverse causation is a documented problem: a viable pregnancy causes nausea and coffee aversion, so surviving pregnancies look like low-caffeine pregnancies regardless. One 2008 study found caffeine looked harmful only among women whose loss preceded the interview, and a Mendelian randomisation study across 49,996 miscarriage cases found no association.

How much caffeine is in a cup of coffee?

Far more variable than most people assume. A home-brewed 8 oz filter coffee is around 95 mg, but the FDA's own range for 12 oz brewed coffee is 113 to 247 mg. Measured retail samples are wider still: espressos have ranged from 25 to 214 mg in one 97-sample analysis, with a single shot measured at 322 mg in another, and the same specialty coffee ordered from the same outlet on six consecutive days measured between 259 and 564 mg. Counting cups is not a reliable way to stay under a threshold.

Does my husband need to stop drinking coffee before IVF?

No guideline asks him to. ASRM states that caffeine consumption has no effect on semen parameters in men, and the largest meta-analysis found men's caffeine and live birth at odds ratio 0.98. In 2,554 young Danish men, intake up to 800 mg a day showed no association with semen quality. A small study of 80 men found more sperm DNA double-strand breaks above three cups a day, but that is a laboratory marker, and the systematic review of 19,967 men calls the male literature inconsistent and inconclusive.

Are energy drinks and cola the same as coffee for fertility?

Probably, though the evidence is not settled. Four independent datasets separate coffee from soft drinks and find the signal on the soft-drink side: sugared soda drinkers undergoing IVF had 12 to 16 percentage points lower live birth while coffee showed nothing, and heavy cola drinkers among 2,554 men had lower sperm concentrations while consuming under 140 mg of caffeine a day, meaning caffeine cannot be the mechanism. The honest counterweight is that the largest and most recent study — 612 women across 1,572 cycles — found neither sugar-sweetened nor artificially sweetened beverages related to live birth.

---

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. Your clinic may have its own protocol; follow it and raise any questions with your medical team.