The Fertility Cliff at 35 Is Based on Records From the 1700s

Laura Erlich · October 26, 2026

The statistic that one in three women aged 35 to 39 will not conceive within a year of trying comes from a 2004 modeling paper in Human Reproduction that estimates natural fertility based on historical populations—French birth records from 1670 to 1830. It is not an observed figure from women trying to conceive today.

Modern prospective studies that follow contemporary women trying to conceive find something different: roughly 78 to 82 percent of women in their late thirties conceive within a year. Fertility does decline with age, and the decline is real and worth acting on. It is not a cliff, and nothing happens on your thirty-fifth birthday.

I'm Laura Erlich, a doctor of acupuncture and Chinese herbal medicine, and certified in integrative reproductive medicine. Twenty-five years in practice means I've watched a lot of women get counseled with that statistic, and I've watched what it does to the decisions they make afterward.

Here is where the number came from, and what the evidence actually supports.

Where the statistic comes from

The source is Henri Leridon's 2004 paper in Human Reproduction, Can assisted reproduction technology compensate for the natural decline in fertility with age?

It is a careful paper, and Leridon was clear about what he was doing. It is a Monte Carlo simulation — a computer model combining monthly conception probability, miscarriage risk, and the age-dependent probability of becoming permanently sterile. Under natural conditions, the model estimated that about 75% of women starting at age 30 would have a live birth within a year, about 66% starting at 35, and about 44% starting at 40.

Subtract 66 from 100, and you have your one in three.

Two things about that number are worth knowing before anyone quotes it at you.

It is modeled, not observed. Nobody followed a group of 35-year-olds and counted. The figure comes from a simulation, and a simulation is only as good as the fertility parameters you feed it. Those parameters came from historical natural-fertility populations — which is the whole point of using them, because if you want to know how often women conceived with no way to prevent it, you need a population that had no way to prevent it. That is sound demography. It is not a description of a woman alive now.

It is a live-birth figure, not a conception figure. Leridon modeled live births within a year, which means miscarriage is already subtracted. The statistic as it circulates — "one in three won't get pregnant" — is a harder claim than the paper made.

Why those women are not a comparison group

Consider who the women in those historical records were.

No antibiotics. An untreated pelvic infection could end fertility permanently, and frequently did.

No treatment for anything else either — no thyroid medication, no diabetes management, no surgery for endometriosis or fibroids, no way to correct any of the conditions a modern workup finds and fixes.

Chronic undernutrition across large parts of the population, and generally far higher rates of infectious disease.

And most of them had already been pregnant several times, often starting in their late teens, with no recovery between pregnancies. A thirty-eight-year-old in 1740 had usually had five or six children already.

She is not a control group for a thirty-eight-year-old today with a thyroid panel and a prenatal vitamin.

What modern data shows

There are prospective studies — contemporary women, actively trying, followed forward.

Those are not cliff numbers. An 82% one-year conception rate at 35 to 39 against 86% at 27 to 34 is a decline of four percentage points, not a collapse.

The JAMA 2017 cohort of 750 women aged 30 to 44 points the same way on a related question: low ovarian reserve markers did not predict reduced natural fertility in women with no history of infertility.

One honest caveat I would rather you hear from me. The modern prospective literature on this specific age band is thin — the studies above together include only a few hundred women aged 35 and over. They are better evidence than a simulation built on parish records, but they are not a large evidence base, and the people who ran them say so.

The picture that emerges from the modern literature is a decline. A real one. But a decline is not a cliff.

Where the number 35 came from

Nothing happens on your thirty-fifth birthday. No biological event happens on that date.

Thirty-five became the age of concern because of a risk calculation about a procedure. It is roughly the age at which the risk of certain chromosomal conditions in a pregnancy crossed the procedural risk of amniocentesis — so it was the age at which offering the test started to make sense on balance. That was a sensible clinical threshold for deciding who to offer a diagnostic procedure to.

It was never a statement about whether a woman could conceive.

It escaped into the culture anyway, and it now behaves like a diagnosis.

What does change with age?

I am not here to tell you age doesn't matter. Age matters. It is the single most predictive variable in reproductive medicine, and anyone who tells you otherwise is selling something.

Here is what actually changes.

Egg quality declines. Specifically, the rate of chromosomal abnormalities in eggs rises with age, and it rises meaningfully through the late thirties and forties. That is the real mechanism — not that the eggs run out, but that the ones remaining are more likely to carry errors when they divide.

Miscarriage rates rise. That follows directly from the above; it is not a small thing, and I am not going to soften it.

Time to conception lengthens on average. What might have taken three months at thirty may take eight at thirty-eight. Same eventual outcome for many women, longer road to it.

All of that is true. What is not true is the framing. The framing says: over at thirty-five.

The evidence says: slower, with more attrition, over a longer runway than you have been told.

Why the cliff framing leads to worse decisions

This is the part I actually care about, because it costs people money and cycles.

I have watched women go straight to IVF at thirty-six with no basic workup, because they believed they had no time to investigate. Then the cycle fails — and the actual problem was a thyroid issue, or a male factor nobody checked, and that cycle was never going to work regardless of how good the embryology was.

Panic is not a diagnostic strategy. It is expensive, and it skips steps.

The opposite error is just as real. Some people hear "thirty-five isn't a cliff" and take it as permission to wait another three years. That is not what the evidence says either. The slope is real, and three years is three years.

The distinction worth holding onto is between urgency and panic.

Urgency means you start the workup now. Panic means you skip it.

How Chinese medicine thinks about reproductive aging

Chinese medicine has never treated age as a threshold.

The framework describes reproductive capacity as something that declines in phases across a lifetime — gradually, continuously, and at different rates in different people. Which, as it happens, matches the modern prospective data better than the cliff does.

The tradition has always held that the rate of that decline depends on how a person lives. Sleep. Stress physiology. Nutrition. Chronic illness. Recovery after depletion.

That is not mystical. Those are the same variables modern reproductive medicine lists when it talks about modifiable factors — the tradition simply got there from a different direction and named it differently.

Your chronological age is fixed. The environment your follicles are maturing in is not.

What to do if you are 35 or over

Get the full workup at six months, not twelve. This is also ASRM's position: evaluation is warranted after six months without conception for women aged 35 and older, rather than the twelve months that defines infertility generally. Not because you are falling off a cliff — because at this stage information is worth more than time, and a workup takes weeks.

Get your male partner tested at the same time. Same appointment if you can. Male factor contributes in roughly half of infertility cases and a semen analysis is the cheapest, fastest test in the whole workup. It is also the one most often skipped.

Ask what population a statistic came from. When a number gets quoted at you — by anyone, including me — that question is not rude. It is the single most useful thing you can ask in a medical appointment. If someone uses a figure to rush you, you are entitled to know whether it describes women like you or women whose records were kept in a parish ledger.

Common questions

Is there really a fertility cliff at 35?

No. Fertility declines gradually, then steepens through the late thirties and forties. The number 35 comes from a threshold for offering amniocentesis, not from a change in fertility on that birthday.

Where does the "one in three women over 35" statistic come from?

A 2004 Monte Carlo modeling study in Human Reproduction whose natural-fertility parameters derive from historical populations, including French birth records from 1670 to 1830. It is a modeled live-birth figure, not an observed conception rate in contemporary women.

What percentage of women aged 35 to 39 conceive within a year?

In modern prospective studies, roughly 78 to 82 percent, compared with 84 to 86 percent for women in their twenties and early thirties.

So does age not matter?

Age matters a great deal — it is the strongest predictor in reproductive medicine. Egg quality declines, chromosomal abnormality rates rise, and miscarriage rates rise with it. The decline is real; the cliff is not.

How long should I try before seeing someone?

Six months if you are 35 or older, twelve if you are younger. Earlier if you have irregular cycles, known endometriosis, prior pelvic surgery or infection, or a partner with a known issue.

Does a low AMH mean I have less time?

Not in the way it is usually presented. AMH estimates follicle quantity, not egg quality, and it poorly predicts natural conception.

Sources

Leridon H. Can assisted reproduction technology compensate for the natural decline in fertility with age? A model assessment. Hum Reprod. 2004;19(7):1548–1553.

Dunson DB, Baird DD, Colombo B. Increased infertility with age in men and women. Obstet Gynecol. 2004;103(1):51–56.

Steiner AZ, Pritchard D, Stanczyk FZ, et al. Association Between Biomarkers of Ovarian Reserve and Infertility Among Older Women of Reproductive Age. JAMA. 2017;318(14):1367–1376.

ASRM Practice Committee. Optimizing natural fertility: a committee opinion.

The Rothman Danish cohort figures (2,820 women, Fertility and Sterility) are cited here as reported by the UNC Time to Conceive study group; the primary paper has not been read directly for this piece.