hCG Doubling Time, and What a Single Number Cannot Tell You

The formula, and where it comes from

Exponential growth means a quantity multiplies by the same factor in equal intervals of time. Write it as B = A × 2^(Δt ÷ T), where T is the time it takes to double. To recover T from two measurements, take natural logarithms of both sides: ln(B) − ln(A) = (Δt ÷ T) × ln(2), so T = ln(2) × Δt ÷ ln(B ÷ A). Nothing in that derivation is specific to pregnancy; it is the same algebra used for half-lives, compound interest and bacterial growth, run in the direction of increase rather than decay.

The formula has one property worth noticing before you use it: it depends only on the ratio of the two values, never on their size. A rise from 120 to 290 over two days is a ratio of 2.42 and a doubling time of 1.57 days; a rise from 12,000 to 29,000 over two days is the same ratio and the same 1.57 days. That is what makes doubling time a fair comparison across different stages, and it is also why the absolute increase is a misleading thing to look at. An increase of 200 mIU/mL over two days is a perfect doubling from a starting value of 200 and a 3.3% rise from a starting value of 6,000 — the same arrow on a chart, a doubling time of two days in one case and forty-two in the other.

The expected rate slows as the level rises

Barnhart and colleagues in 2004 followed 287 women who had pain or bleeding and a non-diagnostic scan but turned out to have a viable intrauterine pregnancy, collecting 861 hCG measurements. The median rise was 1.50-fold after one day, 2.24-fold after two and 5.00-fold after four — three figures that are mutually consistent and correspond to doubling times of 1.71, 1.72 and 1.72 days. That is the middle of the distribution, not a requirement. The same study found the slowest rise seen in a pregnancy that proved viable to be 24% at one day and 53% at two, which are doubling times of about 3.2 days.

The 2016 study from the same group added the piece that is usually missing online: the acceptable minimum depends on where you are starting from. Below 1,500 mIU/mL the slowest 48-hour rise compatible with a viable intrauterine pregnancy was 49%; between 1,500 and 3,000 it was 40%; above 3,000 it was 33%. Converted with the formula, those are doubling times of 3.48, 4.12 and 4.86 days. So a doubling time of four days is outside the expected range at 200 mIU/mL and comfortably inside it at 6,000. Any chart, calculator or forum post that quotes one target for all levels is quoting a figure that stopped being current two decades ago, and the direction of the error matters: it makes normal pregnancies look abnormal at high levels.

Why one value on its own says nothing

Take a widely reproduced laboratory reference table, the one published by UCSF Health for quantitative hCG by week from the last menstrual period. Week 3 runs from 5 to 72 mIU/mL, a fourteenfold span. Week 4 runs from 10 to 708, a seventyfold span. Week 5 runs from 217 to 8,245, thirty-eightfold. Week 6 runs from 152 to 32,177 — a range of more than two hundredfold, spanning better than seven doublings. These are not the extremes of pathology; they are the normal reference intervals for that week.

The practical consequence is that the intervals overlap heavily, so a single value cannot even place you in a week. A result of 300 mIU/mL falls inside the week 4 interval, inside the week 5 interval and inside the week 6 interval simultaneously. A result of 5,000 fits weeks 5, 6 and 7. Reading a number against a chart therefore tells you almost nothing about whether your dates are right or whether the pregnancy is progressing — the chart is too wide to distinguish those cases. Two values taken 48 hours apart in the same laboratory, on the other hand, cancel most of that variation, because the ratio does not care what the starting level was.

Why the rate matters more than the value

Because the ratio removes the starting level, two measurements say something a single one cannot. That is why blood tests in early pregnancy are usually ordered in pairs 48 hours apart, at roughly the same time of day and preferably in the same laboratory: different assays are calibrated differently, and mixing them introduces a difference that has nothing to do with the pregnancy. Where hCG stops being the most useful test is once the level is high enough for an ultrasound to see something directly. ACOG's Practice Bulletin 193 discusses the discriminatory level — the hCG above which an intrauterine gestational sac should be visible — and cautions that if one is used at all it should be set conservatively high, as high as 3,500 mIU/mL, precisely because acting on a lower threshold risks intervening in a pregnancy that is simply early.

That caution is the theme of this whole literature. The 2004 study concluded explicitly that the minimum rise in a viable pregnancy is slower than had previously been assumed, and that intervention should therefore be more conservative rather than less. The numbers were published to slow clinicians down, not to give anyone a pass or fail line. A rate below the expected minimum raises a question — about the location of the pregnancy, about viability, about the dates — and the answer to that question comes from a scan and an examination, not from arithmetic.

What a slow rise or a plateau can mean

A rise slower than the expected minimum is a reason to be seen, not a diagnosis. It is compatible with several situations that look identical on paper and different on a scan: a pregnancy implanted outside the uterus, an early pregnancy loss in progress, dates that are earlier than assumed, or a viable intrauterine pregnancy at the slow end of the distribution — which the Barnhart data show does exist, since the studies were built from pregnancies that turned out viable. A level that plateaus or falls has its own set of explanations, and the same rule applies: which one it is depends on findings the blood test cannot supply.

Two practical points that will save you distress. Do not compare your numbers with someone else's on a forum: given a week 6 reference range spanning 152 to 32,177 mIU/mL, two people at the same gestational age with entirely normal pregnancies can differ by a factor of two hundred, so the comparison carries no information. And do not calculate a doubling time across results from two different laboratories or two different assays, because part of the difference you compute will be the machines rather than the pregnancy. If you are worried, the useful action is a phone call to your clinic, not another calculation.

Frequently asked questions

What counts as a normal doubling time?

There is no single figure, and that is the point of this article. In Barnhart's 2004 series of pregnancies that proved viable the median rise corresponded to a doubling time of about 1.7 days, but the slowest rise seen was 53% over two days, a doubling time of about 3.3 days. The 2016 update made the minimum depend on the starting level: 49% over 48 hours below 1,500 mIU/mL, 40% between 1,500 and 3,000, 33% above 3,000 — doubling times of 3.5, 4.1 and 4.9 days. Any figure you compute has to be read against your own level, your dates and, once it is possible, a scan.

My hCG rose 55% in 48 hours. Is that bad?

It is a number, and on its own it is not an answer. A 55% rise over 48 hours is a doubling time of about 3.2 days. Against Barnhart's published minimums, that clears the 49% floor for a starting level below 1,500 mIU/mL, clears the 40% floor between 1,500 and 3,000, and clears the 33% floor above 3,000. What it cannot tell you is where the pregnancy is implanted or whether the dates are right, and those are the questions that determine what happens next. Send the result to your clinic and ask them, rather than deciding from a threshold.

My value looks low for my week on the charts I found. Should I worry?

The charts are far too wide to support that conclusion. In the UCSF Health reference table, week 5 from the last period spans 217 to 8,245 mIU/mL and week 6 spans 152 to 32,177 — intervals that overlap so heavily that a value of 300 is a normal week 4, a normal week 5 and a normal week 6 result at the same time. Being at the bottom of a range that covers two orders of magnitude tells you almost nothing on its own. What has information in it is a second value 48 hours later from the same laboratory, and what has more is a scan once you are far enough along for one to be informative.

Why did my clinic stop repeating the blood test?

Usually because the level is now high enough for an ultrasound to answer the question directly. ACOG's Practice Bulletin 193 discusses the hCG discriminatory level, above which an intrauterine gestational sac should normally be visible, and advises setting it conservatively high — as high as 3,500 mIU/mL — so that no one acts on a scan that is merely too early. Once you are past that point, a picture is a better test than another number, and the doubling time stops being the most useful measurement.

Can I compare results from two different laboratories?

Better not. Different immunoassays are calibrated differently and can report meaningfully different values for the same sample, so part of any ratio you compute across laboratories is instrument rather than biology — and since the doubling time depends only on that ratio, the error goes straight into your answer. Where possible, have serial tests run by the same laboratory at roughly the same time of day, and give both the values and the exact dates and times to whoever is interpreting them.