Divide the urine protein by the urine creatinine. That is the whole calculation. The reason people still get it wrong is not the division — it is that the units of the answer are decided entirely by the units of the two numbers you put in, and a single missed conversion moves the result by a factor of ten, a hundred or a thousand. This page does the sum three different ways, with real figures, every step shown.
The protein creatinine ratio is urine protein divided by urine creatinine, both measured on the same sample of urine. If your report says protein 96 mg/dL and creatinine 128 mg/dL, the ratio is 96 ÷ 128 = 0.75. That number is 0.75 mg of protein per mg of creatinine, and because a gram is a thousand milligrams, it is also 750 mg/g. Same measurement, two ways of writing it. Everything else on this page exists to stop you getting the second part wrong.
What the ratio is for is a separate question, and a shorter one. A twenty-four hour urine collection is the accurate way to measure how much protein you lose in a day, and it is miserable — a jug in the fridge, a full day of remembering, and one forgotten sample ruins it. The ratio on a single random sample gets you close to the same answer in five minutes, because creatinine is excreted at a fairly steady rate all day and therefore acts as a built-in correction for how dilute the sample happens to be. If you want the interpretation rather than the arithmetic, the normal protein creatinine ratio ranges covers where the cut-offs sit, and what causes a high protein creatinine ratio covers why a result might be raised. This page stays on the maths.
The Waldev creatinine clearance calculator turns a blood creatinine into an estimated filtration rate. New to the molecule itself? What creatinine is explains where it comes from and why it works so well as a reference point.
On this page
The two numbers you need, and where they hide on the report
You need exactly two values, and both must come from the same urine sample. Not from different days, not one from urine and one from blood.
Value 1: urine total protein
Printed as Protein, urine, Total protein, urine or Urine protein (quantitative). Usually in mg/dL in the United States, in mg/L or g/L across most of Europe and the Commonwealth. Typical values on a random sample run from under 10 to several hundred.
Value 2: urine creatinine
Printed as Creatinine, urine or Creatinine, random urine. Usually mg/dL in the US, mmol/L elsewhere. Random-sample values sit anywhere between roughly 20 and 300 mg/dL, or about 2 to 25 mmol/L, depending mostly on how much you have drunk.
The single most consequential thing to check before you divide anything is that the creatinine line says urine. A blood creatinine of 1.1 mg/dL sitting on the same printout looks like a perfectly reasonable number to divide by, and it will hand you an answer roughly a hundred times too large. There is more on that in the mistakes section, because it happens constantly.
Note the units printed next to each value, and write them down before you touch the calculator. Not “96 and 128” — write “96 mg/dL protein, 128 mg/dL creatinine”. Most of the failures on this page trace back to somebody recording two bare numbers and reconstructing the units from memory afterwards. If your report gives urine creatinine in mmol/L and you are used to seeing mg/dL, that is not an error on the lab’s part; it is the same measurement in the units that country uses. Normal urine creatinine levels in mg/dL gives a sense of the range you should expect to see on that line.
One sample, two tests. If protein and creatinine were measured on different specimens, the ratio is meaningless. The whole point of dividing by creatinine is to cancel out the dilution of that particular urine. Two samples, two dilutions, nothing cancels.
The formula, in one line
Here it is, and it does not get more complicated than this at any point.
Protein creatinine ratio = urine protein ÷ urine creatinine
That is the entire calculation. No age term, no sex coefficient, no race variable, no exponent — none of the machinery that makes eGFR equations awkward. If you have ever tried to work through a GFR calculation from creatinine by hand you will appreciate the contrast: that one has powers and multipliers and separate constants by sex, and this one is a single division.
So why does an entire article exist about a division? Because a ratio of two concentrations only means something once you know what units the top and bottom were in. The number 0.75 and the number 750 and the number 84.8 can all be the same patient’s result. Which one you produce depends on nothing but the units you fed in, and the four names for the answer — mg/mg, mg/g, mg/mmol and g/mol — get used interchangeably in conversation by people who are not being careful.
The rule that governs everything: your inputs choose your output
This is the section that fixes the problem, so it is worth reading slowly.
When you divide one concentration by another, the volume units cancel and the mass units survive. That is the whole mechanism. Protein in mg/dL divided by creatinine in mg/dL: the “per dL” on the top cancels the “per dL” on the bottom, and you are left with milligrams of protein per milligram of creatinine. Protein in mg/L divided by creatinine in mmol/L: the “per L” cancels, and you are left with milligrams of protein per millimole of creatinine. Nothing mysterious is happening. The units follow mechanically.
The last row is the trap. Decilitres and litres are different volumes, so nothing cancels and the number that falls out of the calculator is not a ratio of anything. It looks fine. It is exactly ten times wrong. Whenever the top and bottom of your division are expressed per different volumes, one of them must be converted before you divide, and converting the protein is almost always easier than converting the creatinine because it is a straight multiplication by ten.
The four output units you will meet, and how they relate:
| Written as | Means | Where you see it | Relation to mg/g |
|---|---|---|---|
| mg/mg | Milligrams of protein per milligram of creatinine | US labs, often written as a bare decimal such as 0.75 | × 1,000 |
| mg/g | Milligrams of protein per gram of creatinine | US labs and most international guidelines | — |
| mg/mmol | Milligrams of protein per millimole of creatinine | UK, Ireland, Australia, New Zealand, much of Europe | × 8.84 |
| g/mol | Grams of protein per mole of creatinine | Occasionally on European reports | Numerically identical to mg/mmol |
That last row surprises people, so here is the check. One gram per mole is 1,000 mg per 1,000 mmol, which is 1 mg per mmol. The two units are the same size written differently. If your report says 54 g/mol and a guideline says 54 mg/mmol, they agree exactly; no conversion is needed.
Worked example 1: both values in mg/dL
The straightforward case, and the one most American readers will have in front of them. A 54-year-old man with type 2 diabetes has a random urine sample taken at a routine review.
| Line on the report | Value | Units |
|---|---|---|
| Protein, urine | 96 | mg/dL |
| Creatinine, urine | 128 | mg/dL |
Both are per decilitre, so the volumes cancel and no conversion is needed before dividing. Work it through.
Three numbers, one result. If this man’s report prints 0.75 and a guideline he reads online talks in mg/g, he needs 750 to compare like with like, and if he reads a British guideline he needs 84.8. None of those is more correct than the others.
Step five deserves a note, because the direction of the 8.84 catches people. Going from mg/g to mg/mmol you divide by 8.84; going the other way you multiply. The reason is that a gram of creatinine contains 8.84 millimoles of it, so a “per gram” figure is spread over 8.84 times more creatinine than a “per mmol” figure, and it is therefore the bigger number. If your mg/mmol answer ever comes out larger than your mg/g answer, you have used the factor the wrong way round.
One more thing about example 1 before moving on. The value 0.75 mg/mg happens to translate into a rough estimate of about 0.75 g of protein lost per day, and that shortcut is genuinely useful, but it depends on an assumption that does not hold for everybody. That is covered properly further down.
Worked example 2: protein in mg/dL, creatinine in mmol/L
This is the awkward one, and it turns up more often than it should — a lab using conventional units for protein and SI for creatinine, or a patient comparing a result from one country against a reference range from another. A 41-year-old woman with lupus, monitored for kidney involvement.
| Line on the report | Value | Units |
|---|---|---|
| Protein, urine | 45 | mg/dL |
| Creatinine, urine | 6.0 | mmol/L |
Divide these as they stand and you get 7.5, which is not a valid result in any unit system. Decilitres and litres do not cancel. Convert the protein to mg/L first, which means multiplying by ten, because there are ten decilitres in a litre and therefore ten times as many milligrams in a litre as there are in a decilitre.
Now the check, because a mixed-unit calculation is exactly the sort of thing you should verify by a second route. Instead of converting the protein, convert the creatinine and stay in conventional units. Creatinine has a molecular weight of 113.12, so one millimole of it weighs 113.12 mg. A creatinine of 6.0 mmol/L is therefore 6.0 × 113.12 = 678.72 mg/L, which is 67.872 mg/dL. Dividing the original protein by that: 45 ÷ 67.872 = 0.663 mg/mg, which is 663 mg/g.
The two routes agree to three figures. That is the strongest confirmation available that a conversion has been done correctly, and it costs about thirty seconds. Whenever a calculation involves crossing between unit systems, doing it both ways is worth the time — not because the arithmetic is hard, but because a wrong-direction conversion produces a plausible-looking number that nothing else in the result will contradict.
Where 88.4 comes in. The familiar creatinine conversion factor is 88.4, and it converts mg/dL to micromoles per litre: 1 mg/dL = 88.4 µmol/L. Urine creatinine is reported in millimoles per litre, a thousand times bigger, so the direct factor there is 11.312 rather than 88.4. Both come from the same molecular weight of 113.12. Using 88.4 where 11.312 belongs is a thousand-fold error, and it is one of the more common ones.
Worked example 3: SI units throughout
The standard British, Irish, Australian and much-of-Europe presentation. A 67-year-old man being monitored after an episode of glomerulonephritis, with protein reported in grams per litre.
| Line on the report | Value | Units |
|---|---|---|
| Protein, urine | 0.62 | g/L |
| Creatinine, urine | 11.4 | mmol/L |
The volumes match here — both are per litre — so nothing needs converting for the division to work. But the mass unit on top is grams while the conventional output unit is milligrams per millimole, so the protein needs scaling up by a thousand to land on the units everyone quotes.
If you skip step two and divide 0.62 by 11.4 you get 0.054, which is the answer in g/mmol. Not wrong, exactly — it is a correct ratio in an unusual unit — but nobody quotes protein creatinine ratios in grams per millimole, and 0.054 sitting next to a reference range written in mg/mmol is an invitation to conclude that the result is a thousand times better than it is. Convert to mg/L first and the problem never arises.
Notice how differently the three results look in raw form. The headline figures are 0.75, 75 and 54.4, which puts example one at the bottom and example two at the top. Convert them all to mg/g and you get 750, 663 and 481, which puts example one at the top and example three at the bottom. The ranking reverses. Until every figure is in the same unit you are not comparing patients, you are comparing formatting conventions, and that is exactly how a well-meaning comparison against a reference range from another country goes wrong.
Every conversion you might need, with the arithmetic behind it
Four conversions cover essentially every situation. Each one is a single multiplication, and each derives from something concrete rather than being a magic number.
| Converting | Do this | Where the factor comes from | Worked instance |
|---|---|---|---|
| Protein mg/dL → mg/L | × 10 | There are 10 decilitres in a litre | 45 mg/dL × 10 = 450 mg/L |
| Protein g/L → mg/L | × 1,000 | 1 gram = 1,000 milligrams | 0.62 g/L × 1,000 = 620 mg/L |
| Creatinine mmol/L → mg/dL | × 11.312 | Molecular weight 113.12, then per dL not per L | 6.0 mmol/L × 11.312 = 67.87 mg/dL |
| Creatinine mg/dL → µmol/L | × 88.4 | The same 113.12, expressed in micromoles | 1.1 mg/dL × 88.4 = 97.2 µmol/L |
| Ratio mg/mg → mg/g | × 1,000 | 1 gram = 1,000 milligrams | 0.75 × 1,000 = 750 mg/g |
| Ratio mg/mmol → mg/g | × 8.84 | 1 gram of creatinine = 8.84 mmol | 75 × 8.84 = 663 mg/g |
| Ratio mg/g → mg/mmol | ÷ 8.84 | The same factor, reversed | 750 ÷ 8.84 = 84.8 mg/mmol |
| Ratio g/mol → mg/mmol | × 1 (no change) | The two units are identical in size | 54.4 g/mol = 54.4 mg/mmol |
The two factors worth actually memorising are 8.84 and 11.312, and they are the same fact wearing different clothes. Creatinine’s molecular weight is 113.12 grams per mole. Turn that upside down and one gram of creatinine is 1 ÷ 113.12 = 0.00884 moles, which is 8.84 millimoles. So a result of “per gram of creatinine” is a result spread over 8.84 millimoles, and converting between the two is that factor. The 11.312 is the same number scaled for the decilitre: one millimole per litre is 113.12 mg per litre, which is 11.312 mg per decilitre.
You will see 8.85, 8.84 and 8.840 quoted in different places, and occasionally the reciprocal 0.113. They are all the same conversion at different levels of rounding, and the difference between them is far smaller than the biological variation in the measurement itself. A protein creatinine ratio is not a precise instrument. Rounding 8.8402 to 8.84 changes a result of 663 mg/g by less than a milligram, while drinking a litre of water can change the underlying urine concentrations several-fold. Use 8.84 and stop worrying about it.
Converting a reference range instead of your result
Sometimes the tidier move is to leave your own result alone and convert the threshold you are comparing it against. This has one real advantage: you only do it once, and afterwards you can read your own reports natively without any arithmetic at all. Some common landmarks, computed the same way:
| In mg/g | In mg/mmol | In mg/mg |
|---|---|---|
| 150 | 17.0 | 0.15 |
| 300 | 33.9 | 0.30 |
| 500 | 56.6 | 0.50 |
| 1,000 | 113.1 | 1.00 |
| 3,500 | 395.9 | 3.50 |
Those are conversions, not clinical thresholds — where the actual decision points sit depends on the guideline, the reason for testing and the person, and the normal protein creatinine ratio page handles that properly. What the table is for is checking your own arithmetic. If you have converted a result and it lands somewhere sensible against these landmarks, the conversion is probably right; if it lands three orders of magnitude away from all of them, it is not.
The mistakes that put your answer out by ten, a hundred or a thousand
Every error below produces a number that looks entirely reasonable on its own. That is what makes them dangerous. None of them will make your calculator complain.
Dividing mg/dL by mmol/L without converting. Factor of ten. Example 2 done wrong: 45 ÷ 6.0 = 7.5. The true answer is 75 mg/mmol. Seven and a half is a modestly raised figure; seventy-five is a substantial one. Same patient, one missing multiplication, an entirely different conversation with the doctor.
Using micromoles where the report gave millimoles. Factor of a thousand. If the 6.0 mmol/L above is entered as 6,000 µmol/L, the sum becomes 450 ÷ 6,000 = 0.075 instead of 75. This is the classic mistake among people who are familiar with blood creatinine, which is quoted in µmol/L. Urine creatinine is a thousand times more concentrated and is quoted in mmol/L.
Quoting a mg/mg answer as though it were mg/g. Factor of a thousand. Example 2 gives 0.663 mg/mg, which is 663 mg/g. Reporting “0.663 mg/g” describes a result far below any threshold of concern, when the real figure is several times above the usual cut-off. If your ratio is a decimal below 1 and the units say mg/g, something is wrong.
Dividing by the blood creatinine instead of the urine creatinine. Factor of roughly a hundred. Example 1 with the serum value: 96 ÷ 1.1 = 87.3 mg/mg, or 87,273 mg/g, against a true figure of 750. Urine creatinine is typically around a hundred times the blood concentration, because the kidney’s job is to concentrate it. Any ratio in the tens of thousands of mg/g has almost certainly been calculated against a blood value.
Dividing the wrong way round. 128 ÷ 96 = 1.33, where the answer should be 0.75. Protein goes on top, always. The clue that you have inverted it is that both possible answers look plausible when they are near 1, which is exactly when it matters most.
Treating g/L as mg/dL. Example 3’s protein is 0.62 g/L, which is 62 mg/dL. Read as 0.62 mg/dL it is a hundred times too small and produces 0.54 mg/mmol instead of 54.4. Any protein value written with a leading zero and a decimal point is very likely grams per litre.
Multiplying by 8.84 in the wrong direction. mg/mmol figures are always numerically smaller than the same result in mg/g, by a factor of 8.84. If your conversion made the number smaller when you went to mg/g, you divided when you should have multiplied.
Using a dipstick reading as the protein value. A dipstick returns a category — trace, 1+, 2+, 3+ — not a concentration. Those categories are not numbers you can divide with, and the crude mg/dL equivalents sometimes printed alongside them are far too approximate to feed into a ratio. The ratio needs a quantitative protein measurement.
Mixing values from two different samples. Protein from Tuesday’s urine and creatinine from Friday’s cancels nothing. The dilution correction only works within a single specimen.
Run down that list and you will notice that the errors cluster into two families. Volume mismatches cost you a factor of ten. Mass or prefix mismatches cost you a factor of a thousand. If your answer and someone else’s differ by a clean power of ten, the fault is essentially never in the division and essentially always in one of these.
Sanity-checking an answer before you trust it
Do this every time. It takes ten seconds and it catches nearly all of the errors above.
In mg/g, real results run from single figures up to about 15,000 in the most extreme nephrotic cases. In mg/mmol, from under 1 up to a few thousand. In mg/mg, from 0.01 to about 15. A result of 87,000 mg/g or 0.0005 mg/mmol is not a rare finding; it is an arithmetic error.
mg/mg results are small decimals. mg/mmol results are usually in the tens or low hundreds. mg/g results are usually in the hundreds or thousands. A three-digit number labelled mg/mg and a decimal labelled mg/g are both immediately suspicious.
Take your answer, convert it, and ask whether both versions sit in plausible ranges. An error that survives one unit rarely survives two, because the plausible windows are different shapes.
Before dividing, glance at the two raw numbers. On a random sample in conventional units, urine creatinine is usually the larger of the two, so the ratio in mg/mg is usually below 1. If protein exceeds creatinine you should get a ratio above 1, which is possible and significant, but worth a second look at the units first.
Convert the top instead of the bottom, or the bottom instead of the top, and confirm the two agree. This is what caught nothing in example 2 because nothing was wrong, and it is precisely why you do it.
A ratio suggesting nephrotic-range protein loss in someone with no swelling, normal albumin and a normal creatinine is more likely to be a units error than a silent catastrophe. Numbers that disagree violently with everything else usually lose the argument.
That fourth step earns its place. Urine creatinine on a random daytime sample is usually somewhere between 20 and 300 mg/dL, and urine protein in someone without significant kidney disease is usually under 20 mg/dL, so most healthy ratios come out well under 0.2 mg/mg. Knowing that shape in advance means an answer of 4.5 mg/mg makes you check the inputs rather than accept it. Sometimes 4.5 is real and represents heavy protein loss. Often it means the creatinine line was the blood value.
When the lab has already calculated it and your figure disagrees
Most laboratories print the ratio for you. If yours does and your own sum comes out differently, work through this before assuming anyone has made a mistake.
| Size of the discrepancy | Almost certainly | What to check |
|---|---|---|
| Exactly 1,000× | A mg/mg versus mg/g difference, or millimoles read as micromoles | Your figure 0.42 and the lab’s 420 are the same result |
| About 8.8× | One of you is in mg/mmol and the other in mg/g | Multiply the mg/mmol figure by 8.84 and compare again |
| Exactly 10× | A decilitre-versus-litre conversion missed on one side | Re-read the units printed beside the protein value |
| About 100× | A serum creatinine used where a urine creatinine belonged | Confirm the creatinine line says urine |
| A few percent | Rounding, or the lab using unrounded values you cannot see | Nothing. This is normal and clinically irrelevant. |
| Small but not explainable | The lab measured albumin, not total protein | Check whether the report says albumin creatinine ratio |
The last row is the one people miss. Albumin and total protein are different measurements, and the albumin figure is always the smaller of the two, because albumin is only part of the protein in urine. A lab that prints an ACR when you were expecting a PCR has not made an error; it has run a different, more sensitive test. The two are not interchangeable and the thresholds are entirely different.
Rounding deserves a word too. Your result and the lab’s will rarely match to the last digit, because the lab divides the raw analyser output before rounding while you divide numbers that have already been rounded for printing. If the report shows protein 45 mg/dL, the true measured value could be anything from 44.5 to 45.5, which shifts a ratio of 75 mg/mmol by about a unit either way. That is not a discrepancy worth investigating. A difference of a few percent between your figure and the lab’s is expected; a difference of a clean power of ten is a units problem; a difference of 8.8 is a mg/mmol problem.
And if none of the above explains it, the honest answer is to ask. Laboratory staff answer questions like this routinely and would far rather clarify a unit than have someone act on a result they have misread. Nothing on this page substitutes for that conversation, and the ratio is one input among many — your blood creatinine, your estimated filtration rate, your blood pressure and your history all sit alongside it. The NIDDK guidance on kidney disease testing sets out how urine and blood results are read together rather than in isolation.
Why the ratio reads roughly as grams of protein per day
This is the shortcut that makes the ratio so useful, and it is also the one place where a bit of scepticism is warranted.
An average adult excretes somewhere around one gram of creatinine per day. So if a urine sample contains 0.75 mg of protein for every 1 mg of creatinine, and the day’s total creatinine output is about 1 g, then the day’s total protein output is about 0.75 g. The ratio expressed in mg/mg reads directly as grams per twenty-four hours. In mg/g the arithmetic is the same with the decimal point moved: 750 mg/g corresponds to roughly 750 mg, or 0.75 g, of protein per day.
Estimated protein per day (g) ≈ ratio in mg/mg × daily creatinine output (g)
The assumption doing all the work there is that daily creatinine output equals one gram. For a great many people it roughly does. For plenty of others it does not, and the error goes in a predictable direction because creatinine production tracks muscle mass. Take example 1’s ratio of 0.75 mg/mg and run it against a range of realistic daily outputs:
| Person | Approx. daily creatinine output | Estimated protein loss per day |
|---|---|---|
| Small, frail, low muscle mass | 0.7 g | 0.53 g |
| Average adult | 1.0 g | 0.75 g |
| Larger, well-muscled adult | 1.5 g | 1.13 g |
| Very muscular young man | 2.0 g | 1.50 g |
Nearly a threefold spread from an identical ratio. So the shortcut systematically understates protein loss in muscular people and overstates it in frail, elderly or very small ones — which is unfortunate, because frail elderly patients are exactly the group where the estimate gets used most. The figures in that table are approximations, not a lookup you should apply to yourself, and the point is only that the conversion carries real uncertainty. It is the same muscle-mass dependency that makes blood creatinine an imperfect marker of kidney function, described in more detail in what creatinine is and what creatinine in urine means.
None of this stops the ratio being the right test in almost every situation. A twenty-four hour collection has its own errors, and they are usually bigger: a missed void, a forgotten start time, a sample left out of the fridge. An imperfect estimate you actually complete beats a theoretically superior measurement you get wrong. The ratio also allows repeat testing at intervals, which matters more than one-off accuracy when the question is whether a number is rising.
The other assumption worth naming is that a random sample represents the day. It largely does, but not perfectly. Protein excretion in some people rises with upright posture and activity, so an afternoon sample can read higher than an early-morning one. That is why a first-void morning sample is often requested — it removes most of the postural effect and makes serial results more comparable. If you are tracking a ratio over months, collecting at the same time of day matters more than any of the arithmetic on this page.
The same arithmetic, applied to the albumin ratio
Every rule above transfers unchanged to the albumin creatinine ratio. Swap albumin in for total protein and divide exactly as before.
Albumin creatinine ratio = urine albumin ÷ urine creatinine
An instance in conventional units. Urine albumin 3.2 mg/dL, urine creatinine 88 mg/dL. Both per decilitre, so divide directly: 3.2 ÷ 88 = 0.0364 mg/mg, which is 36.4 mg/g, which divided by 8.84 is 4.1 mg/mmol.
The same case in SI units. Urine albumin 32 mg/L, urine creatinine 7.8 mmol/L. Litres cancel: 32 ÷ 7.8 = 4.1 mg/mmol, and multiplying by 8.84 gives 36 mg/g. The two routes land in the same place, as they must.
Two differences from the protein ratio matter in practice. First, albumin concentrations are much smaller than total protein concentrations, so an albumin result in mg/mg is a very small decimal and is almost never quoted that way — mg/g and mg/mmol dominate. Second, and more importantly, the thresholds are completely different and an order of magnitude lower, because albumin is a fraction of total protein and appears in urine earlier in kidney damage. Comparing an ACR against a PCR threshold, or the reverse, produces nonsense. The albumin creatinine ratio calculation covers that test on its own terms.
Read the label before you convert. If a report gives urine albumin in mg/L and urine creatinine in mg/dL, the volumes do not match and the same decilitre trap applies. Convert one side, then divide. The albumin ratio has no special rules — it just has a smaller numerator.
Doing the sum on paper, on a phone, and in a spreadsheet
Three practical ways to run this, in ascending order of how often you will be doing it.
On paper
Write both values with their units on the same line, convert whichever one needs it, draw the division, and write the unit of the answer before you calculate the number. Writing “mg/mmol =” on the page first forces you to think about which conversion belongs where, and that alone eliminates most of the errors in the mistakes section. It feels pedantic. It works.
On a phone calculator
Do the conversion as a separate keystroke rather than combining everything into one long expression. For example 2, key 45 × 10 = to get 450, look at it, confirm it is what you expect, then ÷ 6 = to get 75. Chaining it into 45 × 10 ÷ 6 works arithmetically but gives you nothing to check halfway. Intermediate values are worth seeing.
In a spreadsheet
If you are tracking results over time, a spreadsheet is worth twenty minutes of setup. Put the protein in one column, the creatinine in the next, and the units in a third so you never lose them. The formulae are trivial:
Both in mg/dL, want mg/g: =A2/B2*1000
Protein mg/dL, creatinine mmol/L, want mg/mmol: =A2*10/B2
Protein g/L, creatinine mmol/L, want mg/mmol: =A2*1000/B2
Any mg/mmol result to mg/g: =C2*8.84
Check the first row of a new sheet by hand before trusting the rest. A spreadsheet formula with a misplaced factor is worse than a mental calculation with the same error, because it repeats itself silently down two hundred rows and looks authoritative doing it. Enter 96 and 128 into the first formula and confirm it returns 750; enter 45 and 6.0 into the second and confirm it returns 75. If both check out, the sheet is sound.
A fourth column giving the date is worth adding. The trajectory of a protein creatinine ratio over a year tells you far more than any single value, in the same way that a series of blood creatinine results tells you more than one reading — a point made at greater length in what a normal creatinine level is. A ratio that has been stable at 60 mg/mmol for three years is a different situation from one that has climbed from 15 to 60 in six months, even though today’s number is identical.
Quick-reference conversion card
Everything above, compressed to what fits on a card.
The formula
urine protein ÷ urine creatinine, same sample
Before dividing
Make the volume units match. dL and L do not cancel.
Input conversions
protein mg/dL → mg/L: ×10
protein g/L → mg/L: ×1,000
creatinine mmol/L → mg/dL: ×11.312
Output conversions
mg/mg → mg/g: ×1,000
mg/mmol → mg/g: ×8.84
mg/g → mg/mmol: ÷8.84
g/mol = mg/mmol: no change
Expected shapes
mg/mg: small decimal, usually under 1
mg/mmol: single figures to low hundreds
mg/g: tens to low thousands
Rough daily loss
ratio in mg/mg ≈ grams of protein per day, assuming about 1 g of creatinine excreted daily
Three worked results to check any tool or formula against: 96 and 128 mg/dL gives 0.75 mg/mg and 750 mg/g. 45 mg/dL with 6.0 mmol/L gives 75 mg/mmol and 663 mg/g. 0.62 g/L with 11.4 mmol/L gives 54.4 mg/mmol and 481 mg/g. If a calculator reproduces all three, it is handling units correctly.
Where this number sits among your other results
A protein creatinine ratio on its own answers one question: how much protein is leaking into your urine relative to how concentrated that urine is. It says nothing directly about filtration, and the two can move independently. Someone can have substantial proteinuria with a normal filtration rate, and someone else can have badly reduced filtration with barely any protein in the urine.
That is why the ratio almost never gets ordered alone. It arrives alongside a blood creatinine, an estimated filtration rate derived from it, and usually a set of electrolytes. Creatinine clearance and how clearance compares with GFR cover the filtration side; the ratio covers the leakage side; and it is the combination that describes what the kidney is doing. A high ratio with preserved filtration often means damage to the filter membrane specifically, while a rising ratio alongside falling filtration is a different and more concerning pattern. Neither judgement is one to make from a calculator.
Two other ratios turn up on the same reports and are easy to confuse with this one. The BUN creatinine ratio is a blood test comparing urea to creatinine, used mainly to distinguish reduced blood flow from kidney damage. The albumin creatinine ratio is the urine test described above. Three different ratios, three different purposes, and the only thing they share is creatinine in the denominator. The MedlinePlus overview of creatinine testing is a reasonable place to see how the various creatinine-based tests relate.
One practical consequence. If your ratio has been calculated correctly and comes out raised, the next step is a conversation, not a second calculation. A single raised ratio is frequently repeated before anyone acts on it, because fever, heavy exercise, a urinary infection and even standing for long periods can all lift it temporarily. What matters is whether it stays up.
The ratio tells you about protein leaking out. Filtration is the separate question, and the Waldev creatinine clearance calculator estimates it from your blood creatinine, age, sex and weight. It sits with the rest of the Waldev health calculators.
Related reading across the cluster: normal protein creatinine ratio ranges, what causes a high ratio, what a high urine creatinine means, what high creatinine means, and how clearance is used in drug dosing.
How to calculate protein creatinine ratio: frequently asked questions
How do you calculate the protein creatinine ratio?
Divide the urine protein by the urine creatinine, both measured on the same sample. If the units are mg/dL for both, the answer is in mg of protein per mg of creatinine, and multiplying by 1,000 converts it to mg/g. For example, protein 96 mg/dL divided by creatinine 128 mg/dL gives 0.75 mg/mg, which is 750 mg/g. If the two values use different volume units, such as mg/dL against mmol/L, convert one side before dividing or the answer will be out by a factor of ten.
How do you calculate the urine protein creatinine ratio in mg/mmol?
Put the protein into mg per litre and the creatinine into mmol per litre, then divide. Protein given in mg/dL is multiplied by ten; protein given in g/L is multiplied by a thousand. So 45 mg/dL becomes 450 mg/L, and 450 divided by a creatinine of 6.0 mmol/L gives 75 mg/mmol. A result already in mg/g can be converted instead by dividing it by 8.84. Both routes should agree, and checking one against the other is the fastest way to catch a conversion done in the wrong direction.
What is the formula for the protein to creatinine ratio?
Protein divided by creatinine, and nothing else. There is no age term, no sex coefficient and no weight adjustment, unlike the equations used to estimate filtration rate. The complexity is entirely in the units rather than the formula. What the ratio expresses is how much protein the urine contains relative to how concentrated that urine is, which is why creatinine sits in the denominator: it is excreted at a fairly steady rate, so it acts as an internal marker of dilution and removes the need for a timed collection.
Do I use blood creatinine or urine creatinine?
Urine, always, and from the same sample as the protein. This is the single most common serious error in the calculation. Urine creatinine is roughly a hundred times more concentrated than blood creatinine because concentrating it is exactly what the kidney does, so using a serum value of around 1 mg/dL instead of a urine value of around 100 mg/dL inflates the answer about a hundredfold. If your ratio comes out in the tens of thousands of mg/g, check that line on the report before believing it.
How do I convert mg/mmol to mg/g?
Multiply by 8.84. So 75 mg/mmol is 663 mg/g, and 54.4 mg/mmol is 481 mg/g. Going the other way you divide, so 750 mg/g is 84.8 mg/mmol. The factor comes from creatinine’s molecular weight of 113.12: one gram of creatinine contains 8.84 millimoles. The direction is easy to check without remembering the rule, because the mg/g figure is always the larger of the two. If your conversion made the number smaller on the way to mg/g, you have used the factor backwards.
Why does my answer differ from the lab’s by a factor of 1,000?
Almost always a mg/mg versus mg/g mismatch. A ratio of 0.42 mg/mg and one of 420 mg/g are the same result written two ways, and a report that prints the bare decimal without units invites exactly this confusion. The other thousand-fold trap is entering urine creatinine in micromoles per litre when the report gave millimoles per litre, a habit picked up from blood creatinine, which really is quoted in micromoles. Check which of those two applies before assuming anybody has miscalculated anything.
Can I calculate the ratio from a dipstick result?
No. A dipstick returns a category — negative, trace, 1+, 2+, 3+ — which reflects concentration rather than an actual measured quantity, and there is nothing to divide. The approximate mg/dL equivalents sometimes quoted for each grade are far too crude to feed into a ratio, and they are affected by how dilute the sample is, which is the precise problem the ratio exists to solve. A dipstick is a screening test. If it flags protein, a quantitative laboratory measurement follows, and that is what the calculation uses.
Does the ratio really equal grams of protein per day?
Approximately, and only because an average adult excretes about a gram of creatinine daily. A ratio of 0.75 mg/mg therefore suggests roughly 0.75 g of protein per day. The assumption breaks down at the extremes of muscle mass: someone excreting 2 g of creatinine a day is losing nearer 1.5 g of protein at that same ratio, while a frail person excreting 0.7 g is losing closer to 0.53 g. Treat it as a useful approximation with a threefold spread, not a measurement.
Is the albumin creatinine ratio calculated the same way?
The arithmetic is identical — urine albumin divided by urine creatinine, with the same unit rules and the same 8.84 factor. Urine albumin 3.2 mg/dL over creatinine 88 mg/dL gives 36.4 mg/g, or 4.1 mg/mmol. What differs entirely is interpretation. Albumin is only part of total urinary protein, so the numbers are much smaller and the thresholds are roughly an order of magnitude lower. Comparing an albumin result against a total protein reference range, or the reverse, gives a badly misleading answer.
Does it matter which urine sample I use?
It matters more than the arithmetic does. Both values must come from one specimen, since the whole purpose of dividing by creatinine is to correct for the dilution of that particular sample. Beyond that, a first-void morning sample is usually preferred, because protein excretion rises with upright posture and activity in some people, so an afternoon sample can read higher. If you are tracking a ratio over time, collecting at the same time of day makes serial results genuinely comparable rather than approximately so.
The short version
Urine protein divided by urine creatinine, same sample, and that is the whole formula. The difficulty is entirely in units. Make the volume units match before you divide — decilitres and litres do not cancel, and missing that costs a factor of ten. Then name the units of your answer: mg/dL over mg/dL gives mg/mg, and multiplying by 1,000 gives mg/g; mg/L over mmol/L gives mg/mmol, and multiplying by 8.84 gives mg/g. The three worked cases here are 750 mg/g, 663 mg/g and 481 mg/g, each shown by two routes.
Check every answer against the shape you expect. mg/mg results are small decimals, mg/mmol results run from single figures to low hundreds, mg/g results run from tens to low thousands, and anything outside those windows is a units error rather than a rare finding. Use the urine creatinine, never the blood one. Then take the number to a clinician, because the arithmetic is the easy part. Estimate filtration with the CrCl calculator, and read more in the creatinine blog category, the wider health blog, or the full tool library at waldev.com.
Medical disclaimer: This article explains the arithmetic of a laboratory calculation and is general educational information only. It is not medical advice, cannot interpret your individual result, and must not be used to decide whether to seek care, delay care, or change any treatment or medication. Reference ranges and reporting units vary between laboratories and countries, and a protein creatinine ratio means little without your history, medications, blood results and symptoms alongside it. Always discuss your own results with a doctor or qualified healthcare professional. Seek urgent medical attention for much reduced urine output, new swelling of the legs, face or around the eyes, breathlessness, confusion or persistent vomiting.
MedlinePlus on what creatinine testing measures, how samples are taken, and how blood and urine tests differ. Creatinine test explained →
NIDDK on the tests used to assess kidney disease, and why a urine protein result is read alongside blood work. CKD tests & diagnosis →
The National Kidney Foundation on eGFR, what the ranges mean, and how filtration is graded alongside protein loss. Estimated GFR explained →