IMG_2303This post concerns guidelines just released by the American College of Physicians (ACP) concerning prevention of calcium kidney stones. In the article two specific guidelines are proposed. The first, on fluid management, is covered in another post. Here I discuss their views concerning uses of medication.

The discussion here is in advance of what is already on the site. Whereas I and others have put up more than a few articles about fluids, the issue of medications has not arisen. This is because medications have always been used to alter excretion rates of three atoms or molecules in urine thought to alter risk of calcium kidney stones: calcium, citrate, and uric acid. I have not as yet set the foundation for discussion of their physiology, and therefore how medications might affect stone disease by varying their excretion rates.


“Recommendation 2: ACP recommends pharmacologic monotherapy with a thiazide diuretic, citrate, or allopurinol to prevent recurrent nephrolithiasis in patients with active disease in which increased fluid intake fails to reduce the formation of stones. (Grade: weak recommendation, moderate quality evidence)”


As in the first guideline, the authors offer a kind of codicil:

“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.”

Whereas the bland guideline is inoffensive and predictable to anyone who knows the stone literature, the codicil is right and wrong at the same time: Right because no one has ever compared treatment without measurements of stone composition, or urine and blood chemistries to treatment with them. Wrong because what has not been tested may not be therefore wrong to do.

For example, no one has ever compared dialysis without blood measurements to dialysis with blood measurements, or prednisone treatment of minimal change nephrotic syndrome with vs. without repeated urine testing for albumin.

Because there are few or no trials that compare medical treatment with to treatment without monitoring, one can use the same codicil almost anywhere in medicine.

If one were to take the writing seriously as I do, and as others who read the article may do as well, kidney stone prevention becomes a kind of chaos, where a mixture of mild and serious hazards lurk.


Having as I do an academic soul, nothing will satisfy my rhetoric but to begin with a proper background – albeit circumscribed and excessively brisk – about the medications and their mechanisms in stone prevention.

Supersaturation Rules Crystallization

The primary logic of stone disease is that stones require crystal formation, and that supersaturation is indeed the single physical parameter that best represents the free energy of solid phase formation.

These two statements arise from stone crystallography and the physical chemistry of phase change, which are themselves based on physics. Given that there is only one physics, these statements must apply everywhere.

The Urinary Proteome Modulates Crystallization

Even though we have written here about the power of the urine proteome to de-link supersaturation from crystallization, the proteome has no ability to create crystals on its own. Ultimately, supersaturation is required. As supersaturation is lowered, the crystal forming force diminishes and with it the likelihood of new crystals and therefore stones.

Supersaturations Arise from Concentrations

Supersaturation, in turn, arises from the concentrations in urine of the stone forming salts – calcium oxalate and calcium phosphate in the case of calcium stones. Concentrations of these salts arise, in turn, from the concentrations of calcium, oxalate, and divalent phosphate, which increase salt formation, and citrate which – by binding calcium – reduces salt formation. The concentration of divalent phosphate is determined by that of phosphate and by pH which determines the fraction of phosphate ion in the divalent form.

Concentrations Arise from Proportional Excretions


For calcium, phosphate, oxalate and citrate, concentrations represent the proportions between the excretion rates of calcium, oxalate, phosphate and citrate, and that of water. An increase of urine flow rate will therefore reduce their concentrations and supersaturations at whatever set of excretion rates one specifies. For this reason, urine flow rate is sovereign.


For any given range of urine flows, higher excretions of these atoms and molecules must give rise, on average, to higher concentrations and supersaturations. For this reason excretion rates are sovereign.


Habit and biology govern excretion rates for calcium, oxalate, citrate, phosphate, and likewise urine pH. They vary independently of each other over the day and night in thrall to biology (genetics and past behaviors), food (free will and society), and circumstance (the timing of meals). Habit and circumstance govern urine flow rate, overall and from hour to hour. Therefore, as in England before Henry VIII, Church, and Monarch ruled sovereign, so are the two main forces that drive stone formation ruled by two estates that are, like Church and monarch themselves, caught up in their own webs of influences.


This site has not as yet presented the mechanisms that control calcium, oxalate, citrate or phosphate excretions, or urine pH, as they depend upon whole physiologies which need to be themselves put forth. As I have the intent to write a timely post in advance of proper scholarship, I will not reference the main statements at this time, presuming that most experts know the references and that others will trust me for the moment.


Thiazide diuretic agents reduce urine calcium excretion, and have been used for that purpose for decades as a stone prevention. They prevent stones by altering the proportion between urine calcium loss and urine flow rate, so that on average urine calcium concentration falls. No one has found, thus far, any other way that thiazide can affect kidney stone production.

Alkali salts

Bicarbonate, citrate salts, and in principle other citric acid cycle metabolites create an alkali load. The former is immediate, as bicarbonate neutralizes gastric acid and, when present in excess over acid secretion, is absorbed. Citrate is metabolized as citric acid with uptake of a proton, which leads to new bicarbonate creation in blood. Renal proximal tubule cells can excrete the extra alkali as bicarbonate, raising urine pH, or as citrate which does not alter urine pH. The latter occurs because reabsorption of filtered citrate is reduced by alkali loading. Alkali loading also reduces urine calcium by increasing renal tubule calcium reabsorption. But this is true only for the potassium salts; sodium alkali may not work as well.


It was I who first offered the paradigm that high urine uric acid excretion from excessive purine intake might increase calcium stone formation, and that drugs that reduced uric acid production might reduce calcium stone formation. An RCT of allopurinol showed effectiveness. On the other hand, mechanisms linking uric acid to calcium salt crystallization have not convinced either me or my peers. Likewise epidemiological evidence not only does not link urine uric acid to new onset calcium stones but things run opposite – higher values of urine uric acid associate with lower rates of new stone onset.


The Statement is an Hypothesis of the Field

Given the principles of stone disease as noted above, and also my sketchy but correct precis of both the history of the three medications and their actions, the field has up until now acted as if this hypothesis is correct. For urine calcium lowering by thiazide, and urine citrate increase and urine calcium lowering by alkali salts positive treatment results support the hypothesis. For the case of allopurinol, in fact, the positive treatment result is not at one with known mechanisms through which uric acid lowering in urine might reduce stone formation.

The Medications Do Alter Urine Chemistries

I have said, and say again, that thiazide will lower urine calcium, alkali salts will both increase urine citrate and lower urine calcium, and allopurinol will lower urine uric acid excretion.


The Medications Do Not Always Alter Urine Chemistries

The codicil would have us use medications without knowing stone composition, urine composition, or blood composition. Here, I consider lack of urine and blood measurements.


Urine Sodium

High sodium intake reduces the effectiveness of thiazide in lowering urine calcium excretion. The mechanisms are at least partly related to reduced proximal tubule reabsorption. Instead, potassium wasting may become clinically troublesome and lead to a requirement for potassium supplements. Diet sodium intake cannot be estimated by clinical history; urine sodium is in fact just shy of diet sodium intake.

Serum Sodium

It is true that over years of experience with thiazide I have encountered serious hyponatremia rarely, but I have encountered it. Age, and use of drugs that affect free water clearance can produce this problem in patients urged to drink a lot of fluids and also given thiazide. To use the drug without monitoring serum seems improper.

Urine Calcium

The magnitude of reduction of urine calcium in patients whose urine calcium is in the lower ranges of normal, below 150 or even 100 mg daily, will be low per force. Therefore the drug will be physically futile in lowering supersaturation. Since supersaturation is central to crystallization, failure to lower it is a failure to lower any known mechanism for stone production. In principle, thiazide might act on the urine proteome, but this is pure speculation given an obvious urine calcium and supersaturation effect. To use a futile medication with side effects is not proper clinical behavior.

Individual Variability

For a given dose of thiazide, some patients will show a dramatic and others a minor fall in urine calcium even when baseline urine calcium is high. Therefore wise clinicians begin with small doses and work upward. Sans measurement, one cannot do this.


The codicil must be ignored altogether in respect to thiazide, as low serum potassium is not safe to ignore. Lack of serum monitoring is probable malpractice.


Failure to Use

Even though alkali are given, urine citrate may not rise. One reason is that the alkali is not being taken. Urine measurement of sulfate and ammonia is perfect for monitoring use; alkali will reduce urine ammonia in relation to urine sulfate, and also raise urine potassium.

Inadequate Dose

The effect of alkali is always in relation to net acid production. Patients who have high urine sulfate and ammonia excretions require more alkali to get a citrate response than those with low urine sulfate and ammonia excretions. How do you pick a dose of alkali without measurement?

Undesirable Urine Response

Sometimes, urine bicarbonate may rise in lieu of urine citrate. On the one hand, the medication would then be futile concerning stone prevention. On the other hand, if urine pH is raised urine calcium phosphate supersaturation will rise and raise risk of calcium phosphate stones or conversion from calcium oxalate to calcium phosphate stones. This whole matter is amply discussed in other places but the site does not as yet have sufficient materials to support further discussion.


Patients with diabetes, even mild renal disease, or those taking drugs that interfere with renal potassium excretion may respond to potassium loads with a rise in serum potassium. In general this is not common and when it occurs not life threatening. But monitoring is required for the uncommon but possibly devastating severe hyperkalemic response that leads to death or disability. Potassium loading without serum measurements may possibly represent malpractice.


Dosing For Treatment

In general this drug will almost always lower urine uric acid excretion, but the dosage may well vary between patients. The common 300 mg pill may be unnecessary as opposed to 200 or even 150 mg.

Least Effective Dose

Since side effects are dose related and can be serious one wants to use the least dose that is effective. Failure to ascertain that dose, when possible through testing, is marginal practice quality and in the event of a serious side effect that is known to be dose related may reflect badly on the physician.


The codicil and associated text of the guidelines posits that because “Almost all studies analyzed in the evidence review included only patients with calcium stones, which are the most common stone type.” therefore stone analysis is not helpful. But that is an impossible limitation for a clinician who intends to use medication for stone prevention.

All of the Trials Had Stone Analysis in Them

How can the trials test medications for calcium stones except by having accurate kidney stone analyses to guide them in selection of patients? Without this, patients could have had any kind of stone mixture. A new trial of medications for calcium stones will also depend upon stone analyses. Why exactly does this lead to the codicil remarks that stone analysis results do not affect outcome? It is because such results are always available.

Doctors in Practice Also Need to Have Stone Analyses

Struvite in Stones

Calcium stone formers may have struvite in their stones, at the beginning of their treatment or along the path. Being infected and arising from infection they require specialized medical surgical management. Although sometimes radiographic appearance gives clues to struvite, this is not proven as reliable. Infection of the urine with urea splitting organisms does not prove stones have any struvite in them.

Uric acid in Stones

Since CT scans have replaced the common abdominal flat plate, uric acid stones cannot be differentiated from calcium stones unless an astute radiologist, urologist, or medical specialist measures radiological density. Of the three medications in the guidelines only potassium alkali is a proper treatment for uric acid stones, and that only when the dose is sufficient to raise urine pH above 6. Thiazide and allopurinol are futile. Mixed calcium oxalate uric acid stones are not rare, and for them the three medications may have a role for the calcium oxalate portion but not the uric acid portion. Failure of treatment will be miss-classified as due to inadequate treatment and unless potassium alkali are added – by chance, as it were – failure will lead to more failure.

Brushite or Apatite Stones

Because calcium oxalate stones are much more common, trials of drugs for calcium stone formers are mainly measuring effects for calcium oxalate stone formers. For this reason we actually have no trials at all for calcium phosphate stone formers. In the complete absence of trial data, one would tend to avoid potassium alkali, which will raise urine pH and calcium phosphate supersaturation. Allopurinol is untested and has no rationale apart from a single calcium stone trial which, as I have mentioned, concerns mainly calcium oxalate stone formers. This leaves thiazide as at least rational, and having a probable reason to be effective via lowering of urine calcium excretion rate.

Cystine Stones

Of course these are evident on stone analysis, but the guideline codicil could be misleading. If stone analysis is unnecessary given that most stones are calcium oxalate, how does one know that uncommon stones are not present? Is it that some stones must be analysed? If so, how many, and when? One might think cystine stones and cystinuria will be grossly evident, but this need not be true if no one does urine screening for cystine excess or stone analyses.

Drug Stones

A patient who has formed calcium oxalate stones may go on a drug that creates stones. Perhaps clinical suspicion will lead to a correct diagnosis and avoid futile increases of treatment, perhaps not.


Sensibly, the text of the guideline includes this squib: “Our recommendations do not include patients with suspected hyperparathyroidism or other rare cases.”

But, in the body of the text are these comments: “Results from 1 good-quality (8) and 28 fair-quality trials (9 –35) showed that current evidence is insufficient to conclude that assessing baseline stone composition, blood chemistry, or urine chemistry before initiating pharmacologic or dietary interventions reduces stone recurrence.”

Likewise: “Evidence is insufficient from 1 good-quality (8) and 15 fair-quality trials (9, 11–14, 18–21, 27, 28, 30, 32–34) to conclude that monitoring stone composition, blood chemistry, or urine chemistry once pharmacologic or dietary interventions have been initiated reduces stone recurrence.”

How do we know we are not missing the diagnosis of systemic diseases?

Primary Hyperparathyroidism

The diagnosis is made by finding high serum calcium with unsuppressed serum PTH levels and urine calcium excretion that is not low. The latter gives a clue to familial hypocalciuric hypercalcemia. So serum and urine testing is needed. Of course, having the 24 hour urine sample(s) one knows about hypercalciuria, hypocitraturia, uric acid excretions, pH and supersaturations. Presumably we would not make any other measurements than urine calcium, in order to avoid the extra expense.

Serious hyperoxaluria

Of course primary hyperoxaluria can show up in any calcium oxalate stone former otherwise unannounced but for the urine measurement. Renal failure can be delayed for years. It is not always primary hyperoxaluria. Intestinal malabsorption can be surprisingly free of obvious symptoms. Here and there one encounters patients with remarkable diet patterns and marked elevation of urine oxalate. None will be recognized without testing.

Renal Tubular Acidosis

In general one will find reduced serum bicarbonate and potassium levels, and elevated urine pH and make the diagnosis. Absent serum testing, one presumes that some prior physician will have made the observations. But what is that as a manner of practice? The physician treating stones needs to have the measurements and use them whoever ordered the testing, so blood testing is a necessity. Likewise for urine, as hypercalciuria and low urine citrate do in fact lead to phosphate stones, nephrocalcinosis, and renal disease.


I have exercised myself overly for a purpose, and it is not to rail against what are in fact commonplace maxims about a few reasonably tested drugs or simple statements about fluids for stone formers. Without the mathematical sophistication of the author group any one of us could have come up with the same bland results.

The problem is in the codicils.

Treatment without diagnosis or followup is a radical project. If diagnosis required radiation, vast expense, risk, perhaps one might demur but acknowledge validity. Here we are discussing the rationale to undertake long term prevention care with medications yet avoid simple blood and urine tests of minimal cost. Perhaps a few hundred dollars will do for initial workup and subsequently the same at intervals over the years.

To treat with medications and no testing is radical because it imposes risks of unnecessary failure or even dangerous omissions.

The authors are careful to point out that physicians must decide how they should practice, but there is in that care a bit of what I might call – with no uncivil intent – the lack of complete candor. Insurance payers are intent on cutting costs, and here is an excuse. It is the physicians and their patients who will carry the risks.

The Central Tenet of Evidence Based Medicine

I have said in my prior post that this discipline operates, or seems to, under the hypothesis that what is not proved is not correct. That premise is not part of applied science proper. Applied science takes its objectives from perceived unmet needs. Physicians who practice and find their patients do not do well agitate for better means to take care of them. The patients, too, agitate, as they should.

Evidence based medicine scholars behave as though even when physicians and patients are not dissatisfied, and do not agitate, they – the scholars – have a mandate to test what is being done and reject whatever has not been tested according to some predetermined standard.

The results therefore vary in their usefulness, and sometimes, as in this case, have an off flavor, an oddness about them, a seeming lack of attachment to commonplace clinical realities. They can even have some danger.

It is logical, by the main hypothesis, that one should give out drugs for stones absent either pre-treatment or in-treatment diagnosis or monitoring because neither have proven important in reducing new stones. On the one hand this leads to a kind of nodding absurdity. On the other it leads to false practices as I have noted.

The measurements in question do not appear to influence stone recurrence because they are always done: stone analysis, as an example, so one knows the patients have calcium stones. The trial scientists will have measured serum and urine chemistries in order to qualify subjects for their trials, and many trials have sequential data.

Trials that included testing could be compared retrospectively against trials that seemed to avoid testing – if any can be found, but such retrospective comparisons are hardly rigorous.

Prospective trials with and without testing could be organized, but I would speculate that no IRB would permit such a comparison for reasons of safety and reliable classification of patients. Testing could be required but hidden from the treating physicians in half of the patients; will an IRB permit physicians to practice without seeing the test results?

In fact, will physicians practice at all without testing and monitoring?


With respect for the scientists involved and with regret I believe this publication is untrue and possibly not fully safe.

The idea of medical treatment sans monitoring is contrary to common practice, riddled with possibilities for minor and major harm, and ultimately silly.

If the codicils were not present the guidelines would be essentially trivial, as the three medications have trials everyone knows about and that have been reviewed many times in other places.

With its codicils, this work is unacceptable.



Scroll to Top