IVF Success at 38–40 With Donor Eggs Matches Under-35 Rates: The 50.5% vs 51% Data Point That Reframes the Age Cliff
The Number That Reframes Every Conversation About IVF After 38
A woman aged 38 to 40 pursuing IVF with her own frozen embryos has a 32.7% live birth rate per transfer cycle, according to 2021 CDC ART surveillance data. The same woman, same uterus, same clinic — but using donor eggs — achieves a 51.0% live birth rate.
That 18.3 percentage point gap is not a rounding artifact. It is one of the most actionable signals in the entire fertility dataset, and most couples never see it presented this way.
Here is the part that makes it more interesting: the donor-egg rate for a 38-to-40-year-old (51.0%) is statistically indistinguishable from the live birth rate for a woman under 35 using her own frozen non-donor embryos (50.5%). A half-point difference across a 3-to-5 year age gap. The uterus, in other words, does not meaningfully age between 35 and 40 for purposes of embryo implantation. What changes is exclusively the egg.
This distinction has real financial weight. If the age cliff is fundamentally an egg-quality problem rather than a uterine problem, then the entire decision framework around IVF sequencing — when to add preimplantation genetic testing, when to consider donor eggs, how many retrieval cycles to attempt before pivoting protocol — shifts in ways that are not obvious from how clinics typically present aggregate success statistics.
How the Age Gradient Actually Looks in the Data
Pull the frozen non-donor transfer numbers across age cohorts from the CDC's national ART summary and the decline is steep and consistent:
- Under 35: 50.5% live birth rate per transfer (2021)
- Ages 38–40: 32.7% (2021)
- Ages 41–42: 22.5% (2020)
- Ages 43–44: 13.4% (2018)
The drop from under-35 to 38-40 is 17.8 percentage points. From 38-40 to 41-42 it falls another 10 points. From 41-42 to 43-44 another 9 points. The slope steepens, and it steepens fast.
The conventional narrative around these numbers treats age as the causal variable. The implicit message patients often receive is that the body simply becomes less capable of carrying a pregnancy as a woman moves through her late thirties and early forties. That framing is wrong in a specific and important way.
The donor-egg data breaks it apart. When a 38-to-40-year-old uses a donor egg — which means the egg comes from a woman typically under 30 — the live birth rate jumps from 32.7% back to 51.0%. The uterus has not been upgraded. The endometrium has not been rejuvenated. The hormonal support protocol is essentially the same. The only thing that changed is the source of the genetic material.
That result isolates the variable. The age cliff in IVF is an egg-quality cliff, not a uterine-competence cliff.
Why This Matters for Protocol Sequencing
The practical implication is not simply "use donor eggs if you're over 38." The data is more nuanced than that, and the decision is both medical and deeply personal. What the data does change is how couples should think about information-gathering early in the process.
Most couples who begin IVF in their late thirties start with their own eggs. That is the standard first step, and for many it is the right one — the 32.7% rate is not negligible, and a significant portion of couples in this cohort succeed on their own gametes. But the decision about how many retrieval and transfer cycles to attempt before reassessing the protocol, whether to pursue PGT-A testing on retrieved embryos to screen for aneuploidy before transfer, and at what point donor eggs become a serious option — all of these are more tractable questions if you start with the correct mental model.
The incorrect mental model is: "IVF gets harder as I get older." The correct model is: "IVF with my own eggs gets harder as I get older because egg quality declines, but the embryo-carrying process stays largely intact." The second framing opens up different questions. It makes PGT-A testing more compelling earlier — because aneuploidy rates in retrieved eggs rise sharply after 37, and knowing whether a specific embryo is euploid before transfer changes the expected value of a transfer cycle significantly. SART data consistently shows euploid frozen transfers achieving live birth rates that compress across age groups, because the aneuploidy variable has been removed.
It also changes the calculus on retrieval cycles. A 39-year-old with low ovarian reserve who has produced two retrieved embryos but no confirmed euploid blastocysts faces a different expected-value calculation than the aggregate 32.7% figure implies. The aggregate includes patients who retrieved 8 eggs, banked 4 blastocysts, confirmed 3 euploid — people whose odds at transfer are meaningfully higher than the cohort mean. Understanding where your specific clinical picture sits within the distribution matters more than the headline rate.
You can explore clinic-level and cycle-level data breakdowns at Feralyx to see how these age-stratified rates vary by treatment type and transfer protocol.
The Donor Egg Convergence and What It Tells Us About Uterine Age
The 51.0% vs 50.5% overlap between the 38-40 donor cohort and the under-35 non-donor cohort is the most analytically striking finding in this dataset, and it deserves its own treatment.
Fertility medicine has long understood that egg quality, not uterine receptivity, drives the majority of age-related IVF outcome decline. The ASRM position on age and fertility makes this explicit. But there is a difference between knowing this conceptually and seeing it quantified in cycle-level data.
The near-perfect convergence in the numbers does the following: it establishes that a woman in her late thirties who achieves a high-quality euploid embryo transfer — whether through donor eggs or through her own eggs subjected to rigorous PGT-A selection — should expect implantation odds similar to a younger patient. The uterus at 39 is, from an embryo-acceptance standpoint, nearly as capable as at 32.
This has a flip side that is equally important. When a transfer fails for a woman in her late thirties using her own non-tested embryos, the failure is statistically far more likely to have been driven by embryo chromosomal abnormality than by any failure of the uterine environment. The literature on recurrent implantation failure consistently finds that after two or three failed transfers, a significant proportion of cases resolve once PGT-A-confirmed embryos are used. The uterus was never the problem. It was bearing the downstream consequence of egg quality.
The Cost Dimension the Aggregate Rate Hides
Knowing the egg-quality vs uterine-function distinction is not just biologically interesting — it has direct cost implications that most couples encounter only after spending six figures.
A single IVF retrieval cycle with PGT-A testing typically runs $15,000 to $20,000 out of pocket in most U.S. markets, including the biopsy and genetic analysis costs. A frozen embryo transfer cycle adds another $3,000 to $6,000. A donor egg cycle, depending on whether a fresh or frozen donor bank is used, can run from $25,000 to $45,000 all-in.
The standard objection to donor eggs is cost. But the correct comparison is not "donor egg cycle cost vs. own-egg transfer cycle cost." It is "expected cost per live birth across the realistic range of cycle attempts" — a fundamentally different calculation.
If a 40-year-old has a 22.5% live birth rate per transfer with her own embryos (and that assumes a euploid embryo is available, which is far from guaranteed at that age given retrieval and fertilization attrition), she may need four to five transfers to achieve a live birth at the cohort mean probability. At $5,000 per transfer plus the amortized cost of retrieval cycles to generate embryos, the cumulative cost of success through the own-egg pathway can easily exceed $80,000 to $120,000, with no guarantee of outcome.
A single donor egg cycle at 51% live birth rate has a much better expected-value profile for many patients in this age group, even at the higher nominal cost per cycle. This is not an argument that donor eggs are always the right choice — individual medical circumstances, emotional considerations around genetic connection, and insurance coverage all shift the math. But presenting the decision as "donor eggs are expensive" without showing the probability-adjusted cost comparison is analytically incomplete.
For a more granular look at how treatment pathway costs and success rates interact across different cycle scenarios, Feralyx lets you model expected outcomes by age group, treatment type, and clinic-reported transfer rates.
The Questions Worth Asking Your Clinic Before Cycle One
Most fertility clinics present success rates as a single headline figure, sometimes broken out by age band, rarely stratified by donor vs. non-donor and almost never shown as expected probability across a realistic multi-cycle pathway.
Given what the data shows, there are specific questions that would extract more useful information before a first cycle begins:
What is your clinic's live birth rate per euploid frozen transfer for my age group specifically? This removes the aneuploidy variable and gives you a cleaner picture of implantation competence, which is the thing the clinic's lab and protocol actually influences.
What is your average number of mature eggs retrieved and usable blastocysts produced per retrieval cycle for patients with my AMH and AFC profile? Aggregate success rates are means across a wide distribution. Your starting ovarian reserve determines where on that distribution you are likely to begin.
At what point in your protocol do you typically recommend reconsidering the treatment pathway — whether toward more retrieval cycles, PGT-A testing, or donor eggs? A clinic that never raises this question until after four failed transfers is costing patients money and time that better-structured decision modeling would have avoided.
The 13.4% live birth rate for the 43-44 cohort using their own frozen embryos is a number that deserves directness. It does not mean success is impossible — it means the expected number of cycles to success, and the cumulative cost, are high enough that a rigorous, probabilistic conversation at intake is a service to the patient, not a discouragement.
Reading the Data as a Decision, Not a Statistic
The pattern in the age-stratified IVF data — 50.5% under 35, 32.7% at 38-40 with own eggs, 51.0% at 38-40 with donor eggs, 22.5% at 41-42, 13.4% at 43-44 — is not primarily a medical story. It is a decision architecture story.
Every one of those numbers represents a different expected-value calculation, a different cost trajectory, a different set of tradeoffs between genetic connection, timing, financial exposure, and probability of success. National averages flatten the variation that matters most for individual planning. The donor-egg convergence with under-35 rates is not visible in clinic brochures. The 18-point gap between own and donor eggs at 38-40 is not in the informed-consent summary most patients receive.
The analytical contribution this data makes is specific: the IVF age penalty is localized in the egg, not the uterus. That single inference, derived from comparing the non-donor and donor rate curves, changes what information is worth gathering, when to gather it, and how to weigh protocol decisions that can easily be separated by $30,000 to $50,000 in realized cost.
If you are working through IVF decisions and want to see how these age-stratified success rates map to specific cycle cost scenarios, explore the full data and modeling tools at Feralyx — built specifically for fertility decision intelligence at the individual planning level, not the population average level.
Related Analysis
Other Smart Technology Investments tools that bear on this decision: