IMG_2303The American College of Physicians has published its Clinical Guidelines on dietary and pharmacological management of kidney stones in adultsMy purposes are to place the results of their deliberations in clinical context and also draw some conclusions about research we might want to perform.

Though only two in number, their recommendations cover virtually our entire field of practice. For this reason I thought it best to consider each of the two separately, in different posts.

As a simplification, I accept the statistical analysis as correct. Partly I suspect it is correct. Partly, I am not engaged with the methods of the analysis so much as I am with how we view their results and their conclusions.

The ACP Guidelines can be influential among primary care physicians. As people involved with kidney stone prevention and management we need to understand the effects the guidelines may have on patient care.


I have nothing but appreciation for the complexity of guideline production, and say at the outset that these are fine of their type. As is usual, the guidelines are general, and relate exclusively to simple endpoints of patient benefit – in this case reduction of stones. Likewise, as is usual in ‘evidence based medicine’ their platform assumption is that what has not been proven is not true – a fair way of looking at things, but necessarily a narrow one.

There are two guidelines and a kind of codicil remark about the data bearing upon blood and urine testing before and during treatment. The codicil is important for understanding what the guidelines mean.


‘Recommendation 1: ACP recommends management with increased fluid intake spread throughout the day to achieve at least 2 L of urine per day to prevent recurrent nephrolithiasis. (Grade: weak recommendation, low-quality evidence).’


Certainly, I think it is. At least two liters of urine achieved by fluids spread out over the day is what I desire at a minimum.

In their explanation or expansion of the recommendation, the authors say:

‘People who already drink recommended amounts of liquids and those in whom increased fluid intake is contraindicated should not be directed to increase their fluid intake further.’

I agree with the middle clause: There are some patients who should not raise their fluid intake.

But I do not accept the first clause.


If a patient remains active and has 2 liters of urine, I see no objection to increasing fluids to achieve 3 liters of urine daily, or more. We have provided evidence on urine flow rate and supersaturation that supports increases up to 4 liters of urine a day. 

In what I consider the best study of increased fluids the high fluid group had an average urine volume of 2.6 liters daily, which is closer to 3 than to 2 liters. The guideline writers actually considered this matter in that they recommend ‘at least’ 2 liters of urine volume daily. But in the explanation phrase the point of ‘at least’ is lost in the idea of those who already drink enough.

Consider the implications for a primary care physician caring for a routine calcium oxalate stone former. Two liters of urine are suggested, but stones remain active. Before adding medications, I would increase urine flow further, assuming the supersaturation present was too high. But the recommendation might potentially lead a primary care physician to proceed to medication at 2 liters of urine daily.



As is pointed out elsewhere, insensible losses in stone formers have been measured at 0.9 liters daily in the absence of higher losses from heavy exertion, hot climates or work environments and the like. So the prescription of – say – 2.6 liters of urine daily, which is a reasonable interpretation of the guideline requires at least 3.5 liters of fluid. This is more or less what we came to in our analysis. In summer, or under conditions of high insensible fluid losses the total intake needed for such urine volumes may easily exceed a gallon.


Not part of the fluid recommendations, but under the heading of ‘Inconclusive Areas of Evidence’ is this note:

‘The current evidence from randomized trials is insufficient to evaluate the benefits of knowing the stone composition, urine chemistry, and blood chemistry related to the effectiveness of treatment.’

I am inclined to accept the statement – the codicil –  as true: no one has proven that testing before, after, or during treatment reduces stone formation. But this recommendation by negation, if taken seriously, will make it impossible to achieve the fluid recommendation in a reliable manner.


If we are to increase fluid intake to achieve two liters or more of urine daily we need to know the baseline urine volume. If 24 hour urine testing is not performed, at least for volume, how shall we know the baseline urine volume? All we have is patient reporting.

Patient Reporting of Urine Volume is untested

There are no trials that show patients can self report urine volume accurately enough to discriminate if the volume is two liters a day or less or more. Therefore we cannot determine if more fluids are needed. Essentially, if we accept the platform assumption of evidence based medicine – what has not been proven is not true – then, being untested altogether, patient reports of urine volume are not suitable evidence upon which to base medical practice.

Absent Trials, My Experience Suggests Patients Are Often Inaccurate

I have not done any study of the matter, but express skepticism that self reports of urine volume are accurate enough to rely on for stone prevention. I have met patients whose urine volumes were barely a liter a day but who believed the volume was much higher because they urinated frequently. Some had normal bladders but simply chose to empty them often. Other had prostatic hyperplasia, or female incontinence. I have met patients who detested water and believed their coffee provided several liters of fluid.

These assertions are without force of proof but certainly reflect common situations encountered in practice of stone prevention. They cast doubt on the implicit assumption that patients can know their urine volume well enough to actuate the recommendation without 24 hour urine testing.


Baseline and treatment pose separate issues. Given advice to drink enough to achieve 2 liters of urine daily, patients will have a goal, and try to judge if their volume has increased and if it has achieved the 2 liters. This is really two problems. It is they who have told us the baseline, and who must judge if the volume has increased and if it has achieved a value above two liters. How can patients or physicians know that has happened?



Evidence based medicine holds that what has not been proven is not true.

Patient assessment of urine volume has not been proven accurate in any trials.

Therefore patient assessment of urine volume is not accurate.

Being inaccurate, patient reporting cannot tell us baseline volumes, change in volume, or in-treatment volumes.

Trials or not, the logic of things makes measurement a necessity.


Obtaining a 24 hour urine sample correctly is not simple. Here, I rely on no trials but on the experience of running a 24 hour urine based university owned kidney stone laboratory for 45 years, and being founder of Litholink, now owned by LabCorp.


Somehow a proper contained needs to get to the patient. Proper containers have a scale for measuring the volume of urine that patients can read accurately. Otherwise, the whole container needs to come to a place where a nurse or technician can measure the volume.


With the container must come instructions for how to collect such a urine. The process is not intuitive. For example, people will not know that one begins a collection by emptying the bladder and discarding the urine. Many expect that the start must mean urine is put into the container.

The users of the container cannot be expected to decipher complex instructions. Diagrams are best, and need to be drawn or purchased. The writing must be clear, and simple enough that people without a higher education can fully understand them and act upon them. For women, one needs to provide some kind of aid to collecting.


Samples are Brought to Technicians or Nurses

In my university kidney stone laboratory experience, many patients would drive to the university with their 24 hour urine collection and our technicians would measure the volume. Those measurements are correct as volumes. We have also, in some cases, had 24 hour urine samples brought to the laboratory by local courier services. This involves some expense, and leakage is possible.

Patients Read the Volumes Themselves and Write Them Down

Requires a Trial

This is not a simple task. A lot depends upon the choice of the container. Likewise, the instructions need to be very particular but also easy to use. No laboratory or practice can rely on patient volume reporting without some formal trial in which patient volumes are written down at home and the entire collection brought or mailed in so the value can be checked. The trial is of the container and instructions, mainly.

Should have a Backup

When patient volumes are used, one should have a backup. One common technique is a small vial containing a safe analyte that is added into the container at the beginning and measured in the final urine. The concentration of the analyte is divided into the amount added – which has been weighed out and is therefore known- to get the volume of the sample. The patient measured volume and the volume determined by dilution must match within a reasonable margin – perhaps 10 to 15 percent.

This latter, an analyte dilution estimate of volume, requires that an aliquot of the sample be sent back to the laboratory that can measure the concentration of that analyte. Common carriers like FedEx will transport urine samples but require proper enclosure devices: a sealed plastic bag for the container; a 50 ml or so container with a cap that does not leak; other details I do not need to burden this text with.

The estimate requires an assay system that has quality control and quality assessment programs to be sure it is consistent and accurate. One needs a calculation, a means of reporting, a means of comparing the estimate with patient reported volume, and a means for arbitraging samples with discordance. In other words, the dilution estimate requires a properly equipped laboratory.


Patients – for they are the only people who can know – must somehow note the start and stop times for the collection, and have a form provided to them so the information can be conveyed to the laboratory.


There is no defense against incorrect times except indirect. A common method of cross checking is to measure urine creatinine, ask patients to provide body weight, calculate mg of creatinine per kg of body weight and compare the results to the ranges for men and women. The problem is with the ranges. They are known, but wide. Very fat or very fit people will be at the extremes. Several baseline urines permit better inferences: if the total creatinine excretions per kg body weight are similar the collections are probably reasonable. As urine samples increase in number it becomes easier to know when a collection is incorrect.



I can almost hear some say, ‘Hey, lets keep it simple. Collect urine at home, write down how much and the start and stop times. No problem.’

For all the reasons I have already mentioned I would hesitate to embrace this as a general approach. We still need the right containers or measurement devices, proper collection instructions, time keeping, and correct measurement using all of these items.

It is not that patients are somehow unable to do tasks like this. It is that the task is more unfamiliar than one might think it is.

Who will supervise? Who will assure things are being done right?

The patients are the ones to pay if volumes are wrong, too low, and physicians do not know because the method has faults in it. I would not condone placing such burdens on patients unless we had a trial that proved they can do this kind of work accurately.

Neither would the authors of the guidelines. ‘Evidence Based Medicine’ should be based on evidence, in this case that patients can ‘keep it simple at home’ within a particular practice or health care organization.


To qualify as a source of 24 hour urine volume measurements a practice needs to prove it can measure accurately, and be qualified as a provider. Otherwise, the practice cannot be paid for this laboratory measurement. Even if already licensed as a laboratory, the practice must establish methods that assure competency in this measurement. Perhaps some practices will do this, but perhaps many will not. The point is not tested.


Some but not all commercial and university laboratory vendors have the knowledge, internal control trials, systems, instructions, and methods to do accurate 24 hour urine volumes. Even among those who are approved for the test, quality varies. For example in our own university the main laboratories have containers and instructions but would require collections be brought in or mailed in. They lack a dilution assay.


I do not know, but doubt they can. The containers, instructions, drawings, forms, mailing out, mailing back, measurement, data handling, reporting – all these cost money. Assuming insurance carriers will pay for 24 hour urine volumes – with urine creatinine measurements assumed – how much will they pay compared to how much it costs to produce the testing?


Even if it were true that 24 hour urine testing does not improve medical prevention of kidney stones, the urine volume alone would be an important factor in justifying its use. Fluid management sans measurement is not likely to work and has never been proven to work.

Just as for all other aspects of ‘evidence based medicine’ we need to be careful about interpretation. Unproved does not mean incorrect; it merely means unproved. As an example, has anyone proved that recommended rates of blood testing in chronic kidney disease improve survival or reduce progression to end stage kidney disease?



Do we have enough trials? The ACP people are adept and not terribly impressed with our trials. Our fluid data are classed as weak.

It is true that the first guideline has a self contradiction and is perhaps naive about 24 hour urine volume, but that is beside the point here.

The physical chemistry of stones speaks to high urine flow as an outstanding remedy that should be able to reduce stone formation almost proportional to reduction of supersaturation. But we have not organized trials of this obvious assertion over the past decades and as a result have no evidence for high urine volumes as a stone prevention. However mundane, trials are needed and I think, as one among many, we could do better. High urine volumes, above 3 liters daily, are the likely magic point from supersaturation considerations.

Do we need to use the guideline as a basis for discussion about some new trials. Does NIH need to consider this matter? I am decidedly not a trialist and so absent myself from further details except to say that a step wise trial of higher volumes might surprise us all in efficacy – or not. We know a lot about volume and supersaturation, and about supersaturation and stone forming propensity. One or two definitive trials might mark out the borderland where fluids can do no more and medications are needed.


I would not recommend trying to follow this guideline as it is written, and would not endorse it in my chapters, books, or lectures or on this site. Without a reliable measurement of urine volume it seems a perfect way to mislead patients and ourselves, with consequent excessive treatment failures.

With reliable volumes, I would use the guideline in its exact form: Above 2 liters daily of urine, meaning above 3 liters of fluids daily. But I would be more generous – as much above 2 liters of urine (above 3 liters of fluid daily) as you can do; safety is absolute, efficacy, though not as yet proven, is more than likely.

I say ‘more than likely’ by argument from extremes. Consider 10 liters of urine; supersaturations will be very low in almost all patients, perhaps below 1. Physical chemistry assures us that supersaturation is the only driving force for crystals; decades of crystallography assure us that crystals are what make stones.

Somewhere between more than 2 liters and 10 liters of urine daily one must find perfect safety against stones. My view is to go as far as is practical for a patient and then use medications if that is not enough.


There is magic in fluids where stone formation is concerned. Whereas the guideline is correct, no trials have proven that urine volumes in the high ranges – 3 to 4 liters daily (4 to 5 liters of fluids) – are more effective in stone prevention than volumes in the lower range of 2 to 3 liters a day (3 to 4 liters of fluid daily), lack of proof does not prove something is wrong. When dealing with a treatment that poses no risks to otherwise healthy people, and might obviate need for medications that are reasonably benign and effective but never without some risk, one is wise to get the most out of that risk free treatment.


Scroll to Top